This file was created by the Typo3 extension sevenpack version 0.7.14 --- Timezone: CEST Creation date: 2013-05-23 Creation time: 13-07-58 --- Number of references 41 book 1870 2002 285 In order to recognize an object, the visual system has to solve two basic problems: The first one is how to recognize an object after spatial transformations, i.e. regardless of its orientation, size and position. The second problem refers to the question how we perceive and categorize different objects as members of the same category. A central hypothesis in this book is that both problems have a similar structure, and can be conceptualized within an integrative transformational account, based on concepts from geometry: Recognition after spatial transformations, on one hand, relies on Euclidean transformation processes that are conceptualized as frame transformations (or coordinate transformations). Categorization up to the basic level, on the other hand, can be accounted for by non-Euclidean topological (i.e. potentially space-curving) transformations. Thus, an integrative theory of recognition and categorization is suggested, based on a process-based interpretation of Felix Klein’s Erlanger Program. Zugl.: München, Univ. , Diss., 2002 http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.wvberlin.de/data/inhalt/graf.htm WVB

Berlin, Germany

Biologische Kybernetik Max-Planck-Gesellschaft en 3-932089-91-X grafmMGraf article 6555 Psychological Science 2010 4 20 4 447-454 Three experiments provide evidence that the perceived similarity between two images is systematically affected by the inherent direction of a transformation that links the two. Participants were shown short animations morphing one object into another from the same basic category. They were then asked to make directional similarity judgments (‘‘How similar is object A to object B?’’) for two stationary images drawn from the morph continuum. Across three experiments, similarity ratings for identical comparisons were higher when the reference object, B, had appeared before the comparison object, A, in the preceding morph sequence. This response to dynamic transformational sequences is in accordance with the view that similarity depends on the ease of transformation between object representations and that transformations between objects in categorization and object recognition are psychologically real. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://pss.sagepub.com/content/20/4/447.full.pdf+html Biologische Kybernetik Max-Planck-Gesellschaft en 10.1111/j.1467-9280.2009.02310.x UHahn JClose grafmMGraf article 6240 Experimental Brain Research 2010 1 200 3-4 251-258 Observing an action activates action representations in the motor system. Moreover, the representations of manipulable objects are closely linked to the motor systems at a functional and neuroanatomical level. Here, we investigated whether action observation can facilitate object recognition using an action priming paradigm. As prime stimuli we presented short video movies showing hands performing an action in interaction with an object (where the object itself was always removed from the video). The prime movie was followed by a (briefly presented) target object affording motor interactions that are either similar (congruent condition) or dissimilar (incongruent condition) to the prime action. Participants had to decide whether an object name shown after the target picture corresponds with the picture or not (picture–word matching task). We found superior accuracy for prime–target pairs with congruent as compared to incongruent actions across two experiments. Thus, action observation can facilitate recogni tion of a manipulable object typically involving a similar action. This action priming effect supports the notion that action representations play a functional role in object recognition. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff Research Group Ernst http://springerlink.metapress.com/content/a5ph883317156381/fulltext.pdf Biologische Kybernetik Max-Planck-Gesellschaft en 10.1007/s00221-009-1953-8 helbigHBHelbig jsteinJSteinwender grafmMGraf MKiefer article 4326 NeuroImage 2007 5 36 Supplement 2 T22-T32 There is convincing evidence for a mirror system in humans which simulates actions of conspecifics. One possible purpose of such a simulation system is to support action prediction in real-time. Our goal was to study whether the prediction of actions involves a real-time simulation process. We motion-captured a number of human actions and rendered them as previous termpoint-lightnext term action sequences. Observers perceived brief videos of these actions, followed by an occluder and a static test posture. We independently varied the occluder time and the movement gap (i.e., the time between the endpoint of the action and the test posture). Observers were required to judge whether the test stimulus depicted a continuation of the action in the same depth orientation. Prediction performance was best when occluder time and movement gap corresponded, i.e., when the test posture was a continuation of the sequence that matched the occluder duration (Experiments 1, 2 and 4). This pattern of results was destroyed when the sequences and test images were flipped around the horizontal axis (Experiment 3). Overall, our findings suggest that action prediction involves a simulation process that operates in real-time. This process can break down when the actions are presented under viewing conditions for which observers have little experience. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.sciencedirect.com/science?_ob=MImg&_imagekey=B6WNP-4NC38V3-5-1&_cdi=6968&_user=29041&_orig=search&_coverDate=12%2F31%2F2007&_sk=999639999.8997&view=c&wchp=dGLbVzz-zSkWz&md5=94c254feab3c9dfc6fb2b6d7b0ffcd4c&ie=/sdarticle.pdf Biologische Kybernetik Max-Planck-Gesellschaft en 10.1016/j.neuroimage.2007.03.017 grafmMGraf BReitzner CCorves ACasile gieseMGiese WPrinz article 4769 Behavioral and Brain Sciences 2007 4 30 2 210-211 Somewhat in contrast to their proposal of two separate somatosensory streams, Dijkerman & de Haan (D&dH) propose that tactile recognition involves active manual exploration, and therefore involves parietal cortex. I argue that interactions from perception for action to object recognition can be found also in vision. Furthermore, there is evidence that perception for action and perception for recognition rely on similar processing principles. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://journals.cambridge.org/action/displayFulltext?type=6&fid=1306160&jid=BBS&volumeId=30&issueId=02&aid=1306152&fulltextType=RA&fileId=S0140525X07001495 Biologische Kybernetik Max-Planck-Gesellschaft en 10.1017/S0140525X07001495 grafmMGraf article 2284 Psychological Bulletin 2006 11 132 6 920-945 A basic problem of visual perception is how we recognize objects after spatial transformations. Three central classes of findings have to be accounted for: (1) Recognition performance varies systematically with orientation, size, and position. (2) Recognition latencies are sequentially additive, suggesting analog transformation processes. (3) Orientation and size congruency effects indicate that recognition involves the adjustment of a perceptual coordinate system. While existing models of object recognition are unable to account for all three types of findings, the data can be explained by a transformational model of recognition (TMR), which relies on coordinate transformations and multiple representations. Recognition is achieved by transforming a generic perceptual coordinate system that defines the correspondence between positions specified in memory and positions in the current visual field so that memory and input representations are aligned. TMR is an analog model of recognition, based on analog (image -like) representations and analog transformation processes. TMR is compatible with models in computational neuroscience proposing that object recognition involves coordinate transformations, implemented by neural gain (amplitude) modulation. These gain modulation processes correspond to transformations of receptive fields in upstream cortical areas. Coordinate transformations seem to underlie both object recognition and visuomotor control, and may be regarded as a general processing principle of the visual cortex. TMR discriminates between compensation processes in object recognition and mental imagery; coordinate transformations in recognition are more closely related to visuomotor control than to mental rotation. This framework has several advantages and overcomes arguments that were raised against alignment models of recognition. http://www.kyb.tuebingen.mpg.de/fileadmin/user_upload/files/publications/psych-bull-2006-920_2284[0].pdf http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff Biologische Kybernetik Max-Planck-Gesellschaft en 10.1037/0033-2909.132.6.920 grafmMGraf article 3345 Experimental Brain Research 2006 4 174 2 221-228 It is typically assumed that perception for action and object recognition are subserved by functionally and neuroanatomically distinct processing streams in the brain. However, recent evidence challenges this classical view and suggests an interaction between both visual processing streams. While previous studies showed an influence of object perception on action-related tasks, we investigated whether action representations facilitate visual object recognition. In order to address this question, two briefly displayed masked objects were sequentially presented, either affording congruent or incongruent motor interactions. We found superior naming accuracy for object pairs with congruent as compared to incongruent motor interactions (Experiment 1). This action priming effect indicates that action representations can facilitate object recognition. We further investigated the nature of the representations underlying this action priming effect. The effect was absent when the prime stimulus was presented as a word (Experiment 2). Thus, the action priming effect seems to rely on action representations specified by visual object information. Our findings suggest that processes of object-directed action influence object recognition. http://www.kyb.tuebingen.mpg.de/fileadmin/user_upload/files/publications/HelbigGrafKiefer_2006_EBR_3345[0].pdf http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://springerlink.metapress.com/content/u22g7hg2r856k062/fulltext.pdf Biologische Kybernetik Max-Planck-Gesellschaft en 10.1007/s00221-006-0443-5 helbigHBHelbig grafmMGraf MKiefer article 2543 Psychological Science 2005 3 16 3 214-221 How do we recognize objects after spatial transformations? Recent neurocomputational models proposed that object recognition is based on coordinate transformations which align memory and stimulus representations. If the recognition of disoriented objects is achieved by adjusting a coordinate system (or reference frame) then recognition should be facilitated when the object is preceded by a different object in the same orientation. Two objects were presented in close temporal contiguity in brief masked displays, either in congruent or incongruent picture plane orientations. In two experiments, naming accuracy was higher for congruent orientations. The congruency effect was independent of superordinate category membership (Experiment 1), and was found for objects with different (horizontal or vertical) main axis of elongation (Experiment 2). The results indicate congruency effects for common familiar objects, even for dissimilar shapes. The findings are compatible with models in which object recognition is achi eved by an adjustment of a perceptual coordinate system. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.blackwell-synergy.com/doi/pdf/10.1111/j.0956-7976.2005.00806.x Biologische Kybernetik Max-Planck-Gesellschaft en 10.1111/j.0956-7976.2005.00806.x grafmMGraf kapingdDKaping hhbHHBülthoff article 1328 Behavioral and Brain Sciences 2001 11 24 3 483-484 http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://journals.cambridge.org/action/displayFulltext?type=1&pdftype=1&fid=88734&jid=BBS&volumeId=24&issueId=&aid=88733 Biologische Kybernetik Max-Planck-Gesellschaft en grafmMGraf WXSchneider article 1331 Zeitschrift für Berufs und Wirtschaftspädagogik 1996 Beiheft 13 23-36 http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Biologische Kybernetik Max-Planck-Gesellschaft RStark HGruber grafmMGraf ARenkl HMandl article 1330 Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie 1995 27 289-312 http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Biologische Kybernetik Max-Planck-Gesellschaft RStark grafmMGraf ARenkl HGruber HMandl inproceedings 2728 Digital imaging in media and education. 28th workshop of the Austrian Association for pattern recognition (OAGM/AAPR) 2004 6 245-252 http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff Burger, W. , J. Scharinger Österreichische Computer Gesellschaft

Wien, Austria

Digital Imaging in Media and Education Biologische Kybernetik Max-Planck-Gesellschaft Hagenberg, Austria 28th Workshop of the Austrian Association for Pattern Recognition (OAGM/AAPR) 3-85403-179-3 DSkokaj grafmMGraf dahlCDahl ALeonardis inproceedings 2230 2003 9 390 http://www.kyb.tuebingen.mpg.de/fileadmin/user_upload/files/publications/pdf2230.pdf http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff Schmalhofer, F. , R. Young, G. Katz Erlbaum

Mahwah, NJ, USA

EuroCogSci 03 Biologische Kybernetik Max-Planck-Gesellschaft Osnabrück, Germany European Cognitive Science Conference 2003 0-8058-5005-8 grafmMGraf hhbHHBülthoff inbook 4491 2010 11 27-52 http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.psypress.com/grounding-sociality-9781848728998 Semin, G. R. , G. Echterhoff Psychology Press

New York, NY, USA

Grounding sociality: neurons, mind, and culture Biologische Kybernetik Max-Planck-Gesellschaft en 978-1-84872-899-8 grafmMGraf SSchütz-Bosbach WPrinz inbook 4492 2010 73-101 Categorization is essential for perception and provides an important foundation for higher cognitive functions. In this review, I focus on perceptual aspects of categorization, especially related to object shape. In order to visually categorize an object, the visual system has to solve two basic problems. The first one is how to recognize objects after spatial transformations like rotations and size-scalings. The second problem is how to categorize objects with different shapes as members of the same category. I review the literature related to these two problems against the background of the hierarchy of transformation groups specified in Felix Klein’s Erlanger Programm. The Erlanger Programm provides a general framework for the understanding of object shape, and may allow integrating object recognition and categorization literatures. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.springerlink.com/content/m543xj3618683ml0/ Glatzeder, B. M. , V. Goel, A. Müller Springer

Berlin, Germany

Parmenides book series on thinking Towards a theory of thinking: building blocks for a conceptual framework Biologische Kybernetik Max-Planck-Gesellschaft en 978-3-642-03129-8 10.1007/978-3-642-03129-8_6 grafmMGraf techreport 4782 2002 11 IST-2000-29375 A firm understanding of how the human visual system recognises and categorises objects is important in order to build a successful cognitive vision system. We have reviewed the relevant literature both on visual object recognition and categorisation (chapter 1). Based on this review and the technical annex of this project we have addressed several topics in a series of psychophysical experiments, focusing on structural aspects of recognition memory, object similarity in the context of categorisation, shape transformations in categorisation, the role of context in recognition and categorisation, and the interplay between object motion and shape for categorisation decisions (chapter 2). Based on our psychophysical results we present our view on recognition and categorisation, proposing an integrative framework that serves as a theoretical basis for a computational recognition system grounded in cognitive research (chapter 3). http://www.kyb.tuebingen.mpg.de/fileadmin/user_upload/files/publications/GraSchWalBul2002_[0].pdf http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://fhweb.imvs.ch/ppt/content/pub/psychophysical-results-from-experiments-on-recognition-categorisation Biologische Kybernetik Max-Planck-Gesellschaft Fachhochschule Nordwestschweiz, Information Society Technologies Programme, Cognitive Vision Systems en grafmMGraf aschwanASchwaninger walliCWallraven hhbHHBülthoff poster 4606 2007 4 Brain areas involved in action representation (premotor cortex, posterior parietal cortex) are activated when subjects name pictures of manipulable objects, like tools (e.g., Chao & Martin, 2000). Moreover, behavioral evidence indicates that manipulable objects potentiate possible actions (e.g., Tucker & Ellis, 1998). Here we investigated whether action representations facilitate object recognition. Subjects were sequentially presented with two pictures of artifactual manipulable objects (e.g., tools, musical instruments) and were required to name the objects. In the congruent condition both objects afford similar actions, while affordances are dissimilar in the incongruent condition. Stimulus pairs in both conditions were matched for baseline naming accuracy, word frequency, word length, and for visual and semantic similarity. Naming accuracy was significantly higher in the congruent condition. This action priming effect substantiates our hypothesis that action representations play an important role in the recognition of artifactual manipulable objects. We further conducted an event-related potentials (ERPs) experiment using the same task and stimuli to determine the temporal and spatial pattern of brain activity underlying the action priming. An effect of action congruency on ERPs started at about 100ms after the onset of the target object at fronto-central electrodes close to the motor cortex. This finding demonstrates that the action priming effect in object recognition arises from rapid access to action representations in motor areas even though subjects did not act upon the objects. Altogether, the results suggest that motor knowledge about potential interactions with objects can influence object recognition within the first 100ms of perceptual processing. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.cimec.unitn.it/events/caos/07/poster_session.htm Biologische Kybernetik Max-Planck-Gesellschaft Rovereto, Italy Rovereto Workshop on Concepts, Actions and Objects (CAOs 2007) en helbigHBHelbig grafmMGraf E-JSim MKiefer poster 4419 Journal of Vision 2006 6 6 6 793 Evidence has accumulated for a mirror system in humans which simulates actions of conspecifics (Wilson & Knoblich, 2005). One likely purpose of such a simulation system is to support action prediction. We focused on the time-course of action prediction, investigating whether the prediction of actions involves a real-time simulation process. We motion-captured a number of human actions and rendered them as point light action sequences. In the experiments, we presented brief videos of human actions, followed by an occluder and a static test stimulus. Both the occluder duration (SOA of 100, 400, or 700 ms) and the distance of the test stimulus to the endpoint of the action sequence (corresponding to 100, 400, or 700 ms) were varied independently. Subjects had to judge whether the test stimulus depicted a continuation of the action in the same orientation, or whether the test stimulus was presented in a different orientation in depth as the previous action sequence. Prediction accuracy was best when SOA and distance to the endpoint corresponded, i.e. when the test image was a continuation of the sequence that matched the occluder duration. This pattern of results was destroyed when the sequences and test images were inverted (flipped around the horizontal axis). In this case, performance simply deteriorated with increasing distance to the end of the sequence. Overall, our findings suggest that action prediction involves a real-time simulation process. This process can break down when the actions are presented under viewing conditions for which we have little experience. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.journalofvision.org/6/6/793/ Biologische Kybernetik Max-Planck-Gesellschaft Sarasota, FL, USA 6th Annual Meeting of the Vision Sciences Society (VSS 2006) en 10.1167/6.6.793 grafmMGraf BReitzner gieseMGiese casileACasile WPrinz poster 3239 2005 2 8 126 It has been proposed recently that object recognition relies on coordinate transformations, i.e. on similar processes as visuomotor control [1]. Thus, the two visual streams involved in object recognition and object-directed action may rely on common computational principles, which provides the possibility for interactions between the two streams. Existing behavioral and neurophysiological findings suggest that viewing manipulable objects automatically potentiates possible actions [e.g., 2,3]. We investigated whether action knowledge has a functional role in visual object recognition. More specifically, we used a priming paradigm to test whether objects are recognized better when viewed after another object which affords congruent as compared to incongruent motor interactions. Two grey-scale pictures of artifactual manipulable objects were presented sequentially (tools, kitchen utensils, musical instruments). Subjects were required to name the objects. The stimuli were briefly presented and masked. The presentation time of the second object was adjusted individually in an adjustment phase so that naming accuracy approached 80%. In the congruent condition both objects afforded a similar motor interaction, and dissimilar motor interactions in the incongruent condition. Stimulus pairs in both conditions were matched for baseline naming accuracy, word frequency, word length, as well as visual and semantic similarity. We found that naming accuracy was higher in the congruent than in the incongruent condition (Experiment 1 and 2). This action congruency effect indicates that object naming is facilitated by a previous activation of an appropriate action representation. In two further experiments we investigated the nature of the representations underlying the action congruency effect. The effect was reduced or absent when the prime stimulus was inverted (Experiment 3), and when the prime was presented as a word (Experiment 4). This suggests that the action representations underlying this congruency effect are closer to specific (parameterised) motor representations than to abstract semantic representations. Overall, the findings suggest that the recognition of manipulable objects involves not only visual but also action representations. These are not abstract semantic representations, but are relatively close to motor representations. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.twk.tuebingen.mpg.de/twk05/programm.php Biologische Kybernetik Max-Planck-Gesellschaft Tübingen, Germany 8th Tübingen Perception Conference (TWK 2005) grafmMGraf helbigHBHelbig MKiefer poster 2620 Journal of Vision 2004 8 4 8 511 How is basic level categorization achieved in the human brain? Behavioral studies showed that categorization performance depends systematically on the amount of morph transformation (e.g. Graf, 2002). These results may be due to time-consuming compensation processes (deformable template matching). If deforming (i.e. spatial) compensation processes are involved, then categorization might not only comprise the ventral visual pathway, as generally assumed, but also the dorsal stream. Objects from 25 common basic level categories were generated by morphing between two category members (using 3ds max). In the categorization task, subjects had to decide as fast as possible whether two sequentially presented objects belonged to the same category. The morph distance between category members was varied (event-related design). In a second task, the same observers perceived intact morphing sequences, scrambled sequences, and static presentations of different morphs (block design). In the categorization task, response latencies increased with increasing morph distance between exemplars. Correspondingly, the BOLD signal increased with increasing morph distance in the lateral occipital complex, the superior parietal and the frontal cortices. Control analyses showed that this pattern of activation cannot be reduced to task difficulty, or increasing shape dissimilarity. In the second task we found parietal activation for the contrast between intact vs. scrambled morphing sequences which was close to the dorsal activation in the categorization task, but not identical. The results suggest that basic level categorization relies on a network of ventral, dorsal and frontal areas. The activation within this network depends systematically on the amount of shape transformation. The dorsal activation seems related to compensation processes in parietal cortex, i.e. spatial (deforming) transformation processes. These findings are in accordance with an alignment approach of categorization. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.journalofvision.org/content/4/8/511.abstract Biologische Kybernetik Max-Planck-Gesellschaft Sarasota, FL, USA Fourth Annual Meeting of the Vision Sciences Society (VSS 2004) 10.1167/4.8.511 grafmMGraf dahlCDahl MErb WGrodd hhbHHBülthoff poster 2773 Perception 2004 8 33 ECVP Abstract Supplement 75 Brain areas involved in action representation (pre-motor cortex, posterior parietal cortex) are activated when subjects name pictures of artifactual objects, like tools (eg Chao and Martin, 2000 NeuroImage 12 478 - 484). Moreover, psychophysical evidence indicates that manipulable objects automatically potentiate possible actions (eg Tucker and Ellis, 1998 Journal of Experimental Psychology: Human Perception and Performance 24 830 - 846). We investigated whether knowledge about action affordances is actually involved in object recognition. We examined with a priming paradigm whether objects with congruent affordances are named more accurately than objects with incongruent affordances. Two gray-scale pictures of artifactual manipulable objects were sequentially presented (tools, kitchen utensils, musical instruments). Subjects were required to name the objects. Stimuli were briefly presented and masked. The presentation time of the second object was adjusted individually in an adjustment phase so that accuracy approached 80%. In the congruent condition, both objects afford the same action (eg 'twisting', 'pouring'), while the affordances differ in the incongruent condition. Stimulus pairs in both conditions were matched for baseline accuracy, word frequency, word length, as well as visual and semantic similarity. Naming accuracy was significantly higher in the congruent condition. These findings substantiate our hypothesis that action affordances play an important role in the identification of artifactual manipulable objects. http://www.kyb.tuebingen.mpg.de/fileadmin/user_upload/files/publications/pdf2773.pdf http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.perceptionweb.com/abstract.cgi?id=v040364 Biologische Kybernetik Max-Planck-Gesellschaft Budapest, Hungary 27th European Conference on Visual Perception helbigHBHelbig grafmMGraf MKiefer poster 2619 NeuroImage 2004 6 22 Supplement 1 e946-e947 How is basic level categorization achieved in the human brain? Deforming shape (morphing) transformations are well suited to describe the shape variability of members of common basic level categories (Figure 1). Behavioral experiments showed that categorization performance deteriorates systematically with increasing amount of morph transformation (Dahl, Graf, and Bülthoff, 2003; Graf, 2002). A possible explanation for these findings is that categorization relies on time-consuming compensation processes (deformable template matching). We tested, using fMRI, whether spatial (deforming) compensation processes are involved in categorization, and hypothesized that such processes would additionally activate the dorsal stream. Objects from 25 common basic level categories were generated by morphing between two members of the same category (using 3ds max). Subjects were familiarized with all objects before the experiment. Eleven subjects participated in two tasks, starting with the categorization task. Subjects had to decide as fast as possible whether two sequentially presented objects belonged to the same basic level category. The transformational distance between category members was varied (event-related design). In a second task, the same observers perceived intact morphing sequences, scrambled morphing sequences, and static presentations of different morph exemplars (updating task, block design). fMRI data were acquired on a 3T scanner (Siemens Trio), with 24 slices of 64x64 voxels every two seconds (resolution of 3x3x5 mm3), and whole brain coverage. Behaviorally, we found that in the categorization task, the response latencies for same trials increased with increasing morph distance between two category members. Correspondingly, the fMRI analysis, contrasting long vs. short morph distance, revealed an increasing BOLD signal in LOC (lateral occipital complex) in both hemispheres. Moreover, we found activation in the superior parietal cortices (BA 7) and in the right frontal cortex (BA 44) (Figure 2). Control analyses showed that this pattern of activation cannot be reduced to task difficulty, or increasing dissimilarity between the objects. In the second task we found dorsal activation for the comparison between intact vs. scrambled morphing sequences (and additional bilateral activation in LOC when intact morphing sequences were compared to scrambled and static sequences). The parietal activation, which was located close to the dorsal activation in the categorization task (Figure 3), seems related to the perception of continuously morphing objects. These results suggest that basic level categorization is not limited to the ventral pathway, but rather relies on a network of ventral, dorsal and frontal activation. The activation within this network systematically depends on the amount of shape transformation. The dorsal activation seems related to compensational processes taking place in parietal cortex, i.e. spatial (deforming) transformation processes. These findings are in accordance with an alignment approach of object recognition and categorization. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.sciencedirect.com/science/article/pii/S1053811905700161 Biologische Kybernetik Max-Planck-Gesellschaft Budapest, Hungary 10th Annual Meeting of the Organization for Human Brain Mapping (HBM 2004) 10.1016/S1053-8119(05)70016-1 dahlCDahl grafmMGraf MErb WGrodd hhbHHBülthoff poster 2679 2004 6 http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff Biologische Kybernetik Max-Planck-Gesellschaft Bolzano, Italy The Depictive Space of Perception: A Conference on Visual Thought grafmMGraf poster 2621 NeuroImage 2004 6 22 Supplement 1 e956-e958 Introduction: Visual categorization is a fundamental capability of the visual system. Deforming shape (morphing) transformations are well suited to describe the shape variability of members of common basic level categories (see Figure 1). We investigated the cortical processes underlying basic level categorization, using region of interest (ROI) analyses. Material and Methods: Objects from 25 common basic level categories were generated by morphing between two members of the same category. Eleven subjects participated in three tasks, starting with the categorization task. Subjects had to decide as fast as possible whether two sequentially presented objects belonged to the same basic level category. The transformational distance between category members was varied (event-related design). In a second task, the updating task, the same observers perceived intact morphing sequences, scrambled morphing sequences, and static presentations of different morph exemplars (block design). In a third task [1] we presented intact objects, scrambled objects, or periods of simple fixation. Subjects were required to perform a one-back-task (shape task, block design). fMRI data were acquired on a 3T scanner (Siemens Trio), measuring 24 slices of 64x64 voxels every two seconds (resolution of 3x3x5 mm3 ), covering the whole brain. Based on the shape processing tasks (red in Figure 2), we calculated ROIs. While the contrast ’intact objects - scrambled objects’ revealed the usual LOC region (Figure 2c), a large network of ventral, dorsal and frontal regions was activated when also the fixation period was included (contrast ’intact - scrambled - fixation’) (Figure 2a,b,d). We used these ventral, dorsal and frontal regions as templates for ROI analysis. Within these ROIs, averaged time courses were calculated using the tools of Kalina Christoff (http://www-psych.stanford.edu/~kalina/SPM99/Tools/roi.html). Results: In the categorization task the contrast long vs. short morph distance revealed an increasing BOLD signal in LOC (lateral occipital complex), in the superior parietal cortex (BA 7), and in the frontal cortex (BA 44) (Figure 2 yellow). ROI analyses showed highly similar pattern of results for ventral and dorsal regions (Figure 3). In the updating task we found dorsal activation for the contrast between intact - (scrambled + static) conditions. This activation spot was close to the dorsal activation in the categorization task, but was not identical (Figure 2b green). Nevertheless, an ROI analysis on the categorization data, based on the dorsal activation in the updating task as a ROI, showed similar results as the previous analyses. Discussion: The results confirm that basic level categorization is not limited to the ventral pathway, but rather relies on a network of ventral, dorsal and frontal areas. ROI analyses showed that the activation within this network is systematically dependent on the amount of shape transformation. Similar patterns of activation were found in the ventral and dorsal stream, suggesting that both visual streams are involved in visual categorization. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.sciencedirect.com/science/article/pii/S1053811905700161 Biologische Kybernetik Max-Planck-Gesellschaft Budapest, Hungary 10th Annual Meeting of the Organization for Human Brain Mapping (HBM 2004) 10.1016/S1053-8119(05)70016-1 MErb grafmMGraf dahlCDahl WGrodd hhbHHBülthoff poster 2618 2004 2 7 101 How is basic level categorization achieved in the human brain? Deforming shape (morphing) transformations are well suited to describe the shape variability of members of common basic level categories. Behavioral experiments showed that categorization performance deteriorates systematically with increasing amount of morph transformation [1, 2]. A possible explanation for these ndings is that categorization relies on time-consuming compensation processes (deformable template matching). If spatial compensation processes are involved, then categorization might not only comprise the ventral visual pathway, as generally assumed, but also the dorsal stream. We investigated this question with functional MRI. Objects from 25 common basic level categories were generated by morphing between two members of the same category (using 3ds max). Eleven subjects participated in two tasks, starting with the categorization task. Subjects had to decide as fast as possible whether two sequentially presented objects belonged to the same basic level category. The transformational distance between category members was varied (event-related design). In a second task, the same observers perceived intact morphing sequences, scrambled morphing sequences, and static presentations of different morph exemplars (block design). fMRI data were acquired on a 3T scanner (Siemens Trio), measuring 24 slices of 64x64 voxels every two seconds (resolution of 3x3x5 mm A ). In the categorization task, the response latencies for same trials increased with increasing morph distance between two category members. Correspondingly, the contrast long vs. short morph distance revealed an increasing BOLD signal in LOC (lateral occipital complex). Moreover, activation increased also in the superior parietal cortex (BA 7) and in the frontal cortex (BA 44). Control analyses showed that this pattern of activation cannot be reduced to task difculty, or increasing dissimilarity between the objects. In the second task we found dorsal activation for the comparison between intact vs. scrambled morphing sequences. This activation spot was close to the dorsal activation in the categorization task, but was not identical. The results suggest that basic level categorization is not limited to the ventral pathway, but rather relies on a network of ventral, dorsal and frontal activation. The activation within this network is systematically dependent on the amount of shape transformation. The dorsal activation seems related to compensational processes taking place in parietal cortex, i.e. spatial (deforming) transformation processes. These ndings are in accordance with an alignment approach of object recognition and categorization. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.twk.tuebingen.mpg.de/twk04/index.php Biologische Kybernetik Max-Planck-Gesellschaft Tübingen, Germany 7th Tübingen Perception Conference (TWK 2004) dahlCDahl grafmMGraf MErb zoeZKourtzi WGrodd hhbHHBülthoff poster 2340 2003 11 44 1 If the recognition of disoriented common objects relies on an adjustment of a reference frame, then recognition should be facilitated if the object is preceded by a different object in the same orientation. Two objects from different basic categories were presented in close temporal contiguity in brief masked displays, either in the same or in different picture plane orientations. Subjects had to name both objects. Before the test phase, presentation times for the second object were individually adjusted such that accuracy was at 80%. Line-drawings of common objects were employed. In Experiment 1, objects were selected such that 12 were from biological and 12 from artifact categories. In Experiment 2, 12 objects with a vertical and 12 with a horizontal main axis were used. In both experiments, naming accuracy was higher when the two objects had congruent orientations. This effect was independent of object category, and of the objects’ main axis. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.psychonomic.org/past-meeting.html Biologische Kybernetik Max-Planck-Gesellschaft Vancouver, Canada 44th Annual Meeting of The Psychonomic Society grafmMGraf DKaping hhbHHBülthoff poster 2229 Journal of Vision 2003 10 3 9 506 Previous research suggested that the shape variability of objects from the same basic level category can be conceptualized by transformations which continuously change object shape (topological transformations). Experiments with line drawings (2D outline shapes) demonstrated that categorization latencies and error rates increase with increasing amount of shape transformation (Graf, 2001). We investigated whether these results generalize to more realistic gray-level images rendered from 3D object models. We also studied the effects of combined shape transformations and image-plane rotations on categorization performance. New category members were produced by morphing between objects from the same basic level category. Subjects were required to decide whether two sequentially presented objects belonged to the same basic level category or not. In Experiment 1 the amount of shape transformation was varied, while in Experiment 2 topological distance and image-plane orientation were manipulated. Categorization performance (latencies and accuracy) deteriorated systematically with increased shape transformation, both for upright (Exp. 1) and for rotated (Exp. 2) objects. Furthermore, Exp. 2 showed that categorization latencies increased with increasing amount of orientation change. There was no interaction between shape transformation and object orientation. The results confirm that categorization performance is systematically related to the amount of shape transformation, both for line drawings and gray-level images, as well as for upright and plane rotated objects. In addition, orientation dependency was corroborated with a basic level categorization task. Finally, categorization processes which compensate for shape changes and plane rotations seem to be independent, confirming previous evidence of independent effects for other combinations of spatial transformations (e.g. Lawson et al., 2000). The results support an image-based model of basic level categorization. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.journalofvision.org/content/3/9/506.short Biologische Kybernetik Max-Planck-Gesellschaft Sarasota, FL, USA Third Annual Meeting of the Vision Sciences Society (VSS 2003) 10.1167/3.9.506 grafmMGraf hhbHHBülthoff poster 2417 Perception 2003 9 32 ECVP Abstract Supplement 117 The shape variability of objects from the same basic-level category can be conceptualised by transformations which continuously change object shape. Morphing between members of a basic-level category allows to systematically vary the shape of common objects. We demonstrated that categorisation performance depends on the amount of shape transformation, using a sequential matching task (Graf and Bülthoff, 2003, EuroCogSci, Osnabrück, Germany). We further investigated shape dependence of basic-level categorisation with name matching and rating tasks. We created five category members (1%, 25%, 50%, 75%, and 100% morphs) for twenty-five basic-level categories, based on morphable 3-D object models. Objects were rendered as grey-level images and presented in a canonical orientation. In the speeded name-matching experiment, subjects had to decide whether an object corresponded to a category name. In a second experiment, subjects were required to rate the typicality of the objects, ie they had to rate how well the objects corresponded to their mental image of the category. Both name-matching latencies and typicality ratings depended systematically on the amount of shape transformation. Name-matching latencies were significantly shorter for exemplars that were more in the middle of the morphing sequence. These exemplars were also rated as more typical in experiment 2. The findings confirm the systematic dependence of categorisation performance on the amount of shape transformation. Shapes that are rated as more prototypical are categorised faster. This shape-dependence in basic-level categorisation is reminiscent of the orientation-dependence in recognition, where the canonical perspective is rated as more typical and recognised faster [Palmer et al, 1981, in Attention and Performance IX Eds J Long, A Baddeley (Hillsdale, NJ: Lawrence Erlbaum Associates) pp 135 - 151]. Overall, the results support an extension of the image-based model of recognition (Bülthoff et al, 1995 Cerebral Cortex 3 297 - 260) to basic-level categorisation. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.perceptionweb.com/abstract.cgi?id=v031197 Biologische Kybernetik Max-Planck-Gesellschaft Paris, France 26th European Conference on Visual Perception dahlCDahl grafmMGraf hhbHHBülthoff poster 2226 Perception 2003 9 32 ECVP Abstract Supplement 117 There is plenty of evidence that object recognition is orientation-dependent, but there is still considerable debate why recognition performance depends on orientation. It was proposed that recognition relies on an adjustment of a perceptual reference frame [Graf, 2002 Form, Space and Object. Geometrical Transformations in Object Recognition and Categorization (Berlin: Wissenschaftlicher Verlag)]--so that the recognition of disoriented objects would be facilitated if it is preceded by a different object in the same orientation. These orientation congruency effects were demonstrated previously for letters (Jolicoeur, 1990 Journal of Experimental Psychology: Human Perception and Performance 16 351 - 364) and for novel objects (Gauthier and Tarr, 1997 Perception 26 51 - 73), but not yet for common objects. Two objects from different categories were presented in close temporal contiguity in brief masked displays, either in the same or in different picture plane orientations. Subjects were required to name the objects. The main dependent measure was the number of errors. Line-drawings of twenty-four common objects were employed. Presentation times for the second object were individually adjusted before the test phase such that accuracy was at 80%, when using a different set of objects. Naming accuracy was significantly higher when the orientation of the second object was congruent with the orientation of the previously presented object. Thus, an orientation congruency effect could be demonstrated also with common objects from different basic-level categories. The results suggest that the recognition of disoriented objects involves an adjustment of an abstract frame of reference which facilitates the recognition of different objects in the same orientation. This finding provides further important constraints for any model of object recognition. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.perceptionweb.com/abstract.cgi?id=v031203 Biologische Kybernetik Max-Planck-Gesellschaft Paris, France 26th European Conference on Visual Perception grafmMGraf DKaping hhbHHBülthoff poster 2321 2003 7 http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.unites.uqam.ca/sccog/indexen.html Biologische Kybernetik Max-Planck-Gesellschaft Montréal, Canada UQÀM Summer Institute in Cognitive Sciences 2003 on Categorisation grafmMGraf hhbHHBülthoff poster 2227 Experimentelle Psychologie 2003 3 45 82 Die Formvariabilität von Objekten einer Basiskategorie kann durch topologische (verformende) Transformationen gut beschrieben werden. Experimente mit Linienzeichnungen zeigten, dass sich die Performanz bei der Kategorisierung systematisch verschlechtert mit zunehmendem Umfang der Formtransformation [Graf, Journal of Vision, 1(3), 98a (2001)]. Wir untersuchten, ob sich diese Befunde generalisieren lassen auf Grauwertbilder (basierend auf 3-D Objektmodellen) und auf in der Bildebene rotierte Objekte. Neue Kategoriemitglieder wurden erstellt durch Morphen zwischen Objekten der selben Basiskategorie. Zwei Objekte wurden sequentiell präsentiert und die Probanden mussten entscheiden, ob beide der gleichen Kategorie angehören. Reaktionszeiten und Fehlerraten nahmen systematisch zu mit dem Umfang der Formtransformation, sowohl für die Grauwertbilder als auch für in der Bildebene rotierte Objekte. Auch mit dem Umfang der Rotation stiegen die Reaktionszeiten an. Es zeigte sich keine Interaktion zwischen topologischer Transformation und Rotation. Die Ergebnisse bestätigen und erweitern die früheren Befunde. Sie legen ein bildbasiertes Modell der Kategorisierung auf Basisebene nahe. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff J. Golz, F. Faul, R. Mausfeld Biologische Kybernetik Max-Planck-Gesellschaft Kiel, Germany 45. Tagung Experimentell Arbeitender Psychologen (TeaP 2003) grafmMGraf hhbHHBülthoff poster 2352 2003 2 6 148 It has been shown that basic level categorization performance deteriorates systematically with increasing shape transformation between two sequentially presented objects, providing evidence for an image-based model of categorization. Furthermore, categorization latencies were shown to be sequentially additive, which suggests that categorization relies on analog compensation processes, i.e. on processes which traverse intermediate points on the transformational path between category representation and perceived stimulus (Graf, doctoral dissertation, Wissenschaftlicher Verlag Berlin, 2002). Using a dierent experimental paradigm, we tried to nd converging evidence that categorization is based on analog shape transformations. Category members from common object categories were produced by morphing between two objects from the same basic level category. Subjects were rst familiarized with all category members. In the experiment, three objects were presented sequentially and subjects were required to decide whether all three belonged to the same category. The third object was a morph whose shape was either in between the rst and the second object (INTER), the same as the second object, or morphed beyond the second object (EXTRA). A signicant main eect of condition was found. Reaction times for the INTER condition were faster then for the EXTRA condition. Thus, categorization decisions were faster when the third object was located on the assumed path of transformation between the rst two objects. Since all objects were familiar, the advantage for the INTER condition seems to result from a previous activation of intermediate shapes in the categorization process, suggesting analog shape transformations in categorization. Other image-based accounts seem less appropriate, but cannot be excluded at present. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.twk.tuebingen.mpg.de/twk03/ Biologische Kybernetik Max-Planck-Gesellschaft Tübingen, Germany 6. Tübinger Wahrnehmungskonferenz (TWK 2003) dahlCDahl grafmMGraf hhbHHBülthoff poster 2228 2003 2 6 150 Shape variability of objects from a basic level category usually can be described well with continuous shape transformations (topological transformations). Previous experiments with line drawings showed that categorization performance depends systematically on the amount of shape transformation, both in speeded categorization tasks and rating tasks (Graf, doctoral dissertation, Wissenschaftlicher Verlag Berlin, 2002). We investigated whether this systematic dependency can be replicated under dierent conditions. Gray-level images of new category members were produced by morphing between members of the same basic level category. Two images were presented sequentially and subjects were required to decide whether they belonged to the same category. The amount of shape transformation between members of the same category was varied systematically. Objects were presented either upright (Experiment 1), in dierent orientations in the picture plane (Experiment 2) or in dierent equidistant positions (Experiment 3). In general, categorization performance deteriorated systematically with increased shape transformation. In Experiment 2, categorization performance depended both on the amount of shape change and orientation change, while there was no signicant interaction between orientation and shape. The eect of shape transformation was found despite position changes (Experiment 3). The systematic dependency of categorization performance on the amount of shape transformation was replicated with gray-level images, for upright objects, image-plane rotated, and for translated objects. This indicates that the systematic dependency on shape changes is not simply due to low-level processes. Processes which compensate for shape and for orientation changes seem independent, which is in agreement with previous ndings for other combinations of transformations. Overall, the ndings strongly support an image-based model of categorization. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.twk.tuebingen.mpg.de/twk03/ H.H. Bülthoff, K.R. Gegenfurtner, H.A. Mallot, R. Ulrich, & F.A. Wichmann Biologische Kybernetik Max-Planck-Gesellschaft Tübingen, Germany 6. Tübinger Wahrnehmungskonferenz (TWK 2003) grafmMGraf hhbHHBülthoff poster 2457 Journal of Vision 2002 11 2 7 691 A previous speeded categorization experiment demonstrated that basic level categorization performance deteriorates with increasing amount of topological (warping) transformation between two successively presented category members (Graf, VSS 2001). If topological transformations are involved in basic level categorization, then performance in related tasks should also be influenced by the amount of shape deformation. In particular, the typicality (or representativeness) of category members should be determined by the transformational distance to the category representation. Also, shape similarity should decrease with increasing transformational distance between the objects. These predictions were investigated for 2D outline shapes of objects from 25 common and familiar object categories. In the typicality task, subjects had to rate the typicality of different category members that were produced with a warping algorithm — i.e. had to judge how well the objects fit with their idea of the category. In the similarity task, the amount of topological transformation between two objects from the same basic level category was manipulated, and subjects had to rate the similarity of the objects. The results confirmed the predictions: First, typicality ratings varied in a systematic way with topological shape transformation: A graded category structure was found, and typicality decreased with increasing distance to the most typical exemplar. Second, perceived similarity decreased in a highly significant way with increasing amount of topological transformation. These findings indicate that basic level categorization performance and perceived shape similarity depend on the amount of topological transformation. The results can be accounted for by an alignment model of categorization and similarity which involves deforming transformations. The suggested model can be regarded as an image-based extension to the structural alignment account of similarity (e.g. Markman & Gentner, 1993). http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.journalofvision.org/2/7/691/ Biologische Kybernetik Max-Planck-Gesellschaft Sarasota, FL, USA Second Annual Meeting of the Vision Sciences Society (VSS 2002) 10.1167/2.7.691 grafmMGraf poster 2225 2002 9 43 Verschiedene Mitglieder einer Objektkategorie auf Basisebene (wie z.B. Fisch, Tasse) lassen sich häufig durch relativ einfache verformende Transformationen ineinander überführen. Ein früheres Experiment zeigte, dass bei einer Kategorisierungsaufgabe die Reaktionszeit und Fehlerrate mit zunehmendem Umfang der verformenden Transformation ansteigen (Graf, 2001). Es wurde deshalb untersucht, ob die Typizität (Repräsentativität) und Formähnlichkeit von Kategoriemitgliedern ebenfalls in systematischem Zusammenhang stehen zu verformenden Transformationen. Mit Hilfe eines Warping-Algorithmus wurden auf Basis von je zwei Mitgliedern einer Kategorie zwischenliegende Exemplare erstellt, und so 2-D Silhouetten für Objekte aus 25 bekannten Objektkategorien erzeugt. In der Typizitätsaufgabe mussten die Probanden beurteilen, wie gut die Objekte zu ihrer Vorstellung der Kategorie passen. In der Ähnlichkeitsaufgabe sollte die Ähnlichkeit von je zwei Objekten einer Kategorie beurteilt werden. Die Ergebnisse bestätigten die Vorhersagen: Zum einen variierten die Typizitätsratings auf systematische Weise mit der Formtransformation; die Typizität nahm mit zunehmender Distanz zum typischsten Kategoriemitglied ab. Zum zweiten verringerte sich die wahrgenommene Ähnlichkeit mit zunehmendem Umfang verformender Transformationen. Die Befunde demonstrieren, dass die Kategorisierung und Formähnlichkeit von Objekten vom Umfang verformender Transformationen abhängen. Die Daten können durch ein Alignment-Modell erklärt werden, das verformende Transformationen beinhaltet. Das vorgeschlagene Modell kann als bild-basierte Erweiterung des structural alignment Modells (z.B. Markman, 2001) betrachtet werden. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.dgps.de/aktivitaeten/kongress/2002/ van der Meer, E.; Hagendorf, H.; Beyer, R.; Krüger, F.; Nuthmann, A.; Schulz, S. Biologische Kybernetik Max-Planck-Gesellschaft Berlin, Germany 43. Kongress der Deutschen Gesellschaft für Psychologie grafmMGraf poster 1332 2002 2 5 82 Verschiedene Mitglieder einer basic level Kategorie lassen sich häufig durch eine relativ einfache topologische (verformende) Transformation ineinander überführen (sog. Gummituch- Geometrie). Ein früheres Experiment zeigte, dass bei einer Kategorisierungsaufgabe unter Zeitdruck die Reaktionszeit und Fehlerrate ansteigen mit zunehmendem Umfang der topologischen Transformation von zwei sequentiell präsentierten Kategoriemitgliedern (Graf, TWK 2001). Falls topologische Transformationen bei der basic level Kategorisierung involviert sind, dann sollte die Performanz in verwandten Aufgaben ebenfalls durch den Umfang der Formdeformation beeinflusst sein. Zum einen sollte die Typizität (Repräsentativität) der Kategoriemitglieder bestimmt sein durch die transformationale Distanz zur Kategorierepräsentation. Zum anderen sollte die Formähnlichkeit von zwei Objekten abnehmen mit zunehmender topologischer Transformation zwischen den Objekten. Diese Vorhersagen wurden untersucht für 2D Silhouetten (outline shapes) von Objekten aus 25 bekannten Objektkategorien. In der Typizitätsaufgabe mussten die Vpn die Typizität von verschiedenen Kategoriemitgliedern beurteilen, die mit einem Morphing-Algorithmus erzeugt wurden – d.h. sie sollten beurteilen, wie gut die Objekte zu ihrer Vorstellung der Kategorie passen. In der Ähnlichkeitsaufgabe wurde der Umfang der topologischen Transformation zwischen zwei Objekten der gleichen basic level Kategorie manipuliert, und die Vpn mussten die Ähnlichkeit der Objekte beurteilen. Die Ergebnisse bestätigten die Vorhersagen: Zum einen variierten die Typizitätsratings auf systematische Art und Weise mit der topologischen Formtransformation. Die Ratings zeigten eine Kategoriestruktur, bei der die Typizität abnahm mit zunehmender Distanz zum typischsten Kategoriemitglied. Zum zweiten nahm die wahrgenommene Ähnlichkeit ab mit zunehmendem Umfang topologischer Transformation. Die Befunde demonstrieren, dass die basic level Kategorisierung und die Formähnlichkeit von Objekten vom Umfang topologischer Transformation abhängen. Die Daten können durch ein Alignment-Modell der Kategorisierung und der Ähnlichkeit erklärt werden, das topologische Transformationen beinhaltet. Das vorgeschlagene Modell kann als eine bild-basierte Erweiterung des structural alignment Modells der Ähnlichkeit (z.B. Markman & Gentner, 1993) betrachtet werden. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.twk.tuebingen.mpg.de/twk02/ Biologische Kybernetik Max-Planck-Gesellschaft Tübingen, Germany 5. Tübinger Wahrnehmungskonferenz (TWK 2002) grafmMGraf poster 1334 Journal of Vision 2001 12 1 3 98 Purpose: Several studies suggest that the image-based approach to object recognition can be extended to basic level recognition, but it is not clear yet how basic level recognition is achieved. It seems that the shapes of category members on the basic level can be aligned by rather simple topological (warping) transformations — for most biological and many artifact categories. This study was motivated by the hypothesis that basic level recognition involves analog topological transformation processes — in order to achieve an alignment of stimulus representation and object representation. Two research questions were investigated: (1) Does categorization performance depend on the amount of topological transformation? (2) If systematic effects can be detected, are these effects caused by analog transformation processes? Methods: Images from 16 biological and 14 artifact categories were scanned. On the basis of two members of a category, with morphing software new exemplars were produced as intermediate morphs at specific transformational distances. In every trial two backward masked images were presented sequentially (sequential matching task). The transformational distance between the two images was varied. Subjects had to decide by key press whether both images belonged to the same basic level category or not. Results: The predictions were confirmed by the data: Both RTs and error rates increased with increasing amount of topological transformation between two objects of the same category. Furthermore, RTs are sequentially additive, which is evidence that the topological transformation traverses intermediate points in the transformational path. Conclusions: Taken together, the data indicate that basic level recognition relies on time-consuming and error-prone topological transformation processes, which have an analog nature. The results support a model of basic level recognition that is based on analog (image-like) representations and analog transformation processes. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://journalofvision.org/1/3/98 Biologische Kybernetik Max-Planck-Gesellschaft Sarasota, FL, USA First Annual Meeting of the Vision Sciences Society (VSS 2001) 10.1167/1.3.98 grafmMGraf poster 1333 2001 3 109 Eine Reihe von Befunden legt nahe, den bild-basierten Ansatz der Rekognition auf das Erkennen von Objekten auf der Basisebene der Kategorisierung (basic level) zu erweitern. Es ist allerdings noch nicht geklärt, wie das visuelle System Objekte auf der Basisebene erkennt. Wenn man verschiedene Mitglieder einer Kategorie auf Basisebene vergleicht, dann scheint es, dass die Formen der Kategoriemitglieder durch relativ einfache topologische (formverändernde) Transformationen ineinander überführt werden können. Dies gilt für die meisten biologischen Kategorien und auch für viele Artefakt-Kategorien. Diese nichteuklidischen topologischen Transformationen können durch die Dehnung bzw. Stauchung eines Gummituchs veranschaulicht werden. Die Studie geht von der Hypothese aus, dass Objekterkennung auf der Basisebene durch analoge topologische Transformationsprozesse erfolgt, welche die Stimulusrepräsentation und die Objektrepräsentation aufeinander ausrichten. Es ergaben sich somit die folgenden beiden Fragen: Zeigt sich bei der Kategorisierung eine Abhängigkeit der Performanz vom Umfang topologischer Transformationen? Falls systematische Effekte auftreten, sind diese durch analoge Transformationsprozesse verursacht? Bilder von Objekten aus 16 biologischen und 14 artifiziellen Objektkategorien wurden gescannt. Mit Hilfe von Morphing Software wurden auf Basis von je zwei Objekten einer Kategorie zwischenliegende Objekte erzeugt, die ebenfalls dieser Kategorie angehören. In jedem Trial wurden zwei maskierte Bilder sequenziell präsentiert (sequential matching task). Die transformationale Distanz zwischen beiden Bildern wurde variiert. Die Probanden mussten durch Tastendruck entscheiden, ob beide Bilder der selben Basiskategorie angehören oder nicht. Beide Vorhersagen wurden bestätigt: Zum einen erhöhten sich sowohl Reaktionszeiten als auch Fehlerraten mit zunehmendem Umfang der topologischen Transformation zwischen zwei Mitgliedern der selben Kategorie. Zum anderen waren die Reaktionszeiten sequenziell additiv, was Evidenz dafür ist, dass die topologischen Transformationen zwischenliegende Punkte im transformationalen Pfad durchlaufen. Das Erkennen von Objekten auf Basisebene scheint folglich auf zeitkonsumierenden und fehleranfälligen topologischen Transformationsprozessen zu beruhen, die analoger Natur sind. Die Ergebnisse legen somit ein Modell nahe, das auf analogen (bildhaften) Repräsentationen und analogen Transformationsprozessen basiert. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.twk.tuebingen.mpg.de/twk01/Pobjekt.htm Biologische Kybernetik Max-Planck-Gesellschaft Tübingen, Germany 4. Tübinger Wahrnehmungskonferenz (TWK 2001) grafmMGraf miscellaneous 1329 Gedächtnis und Erinnerung. Ein interdisziplinäres Lexikon 2001 423-424 http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de N. Pethes & J. Ruchatz Biologische Kybernetik Max-Planck-Gesellschaft grafmMGraf conference 4086 Experimentelle Psychologie 2006 3 48 207 Brain areas involved in action representation (premotor cortex, posterior parietal cortex) are activated when subjects name pictures of manipulable objects, like tools (e.g., Chao & Martin, 2000). Moreover, behavioral evidence indicates that manipulable objects potentiate possible actions (e.g., Tucker & Ellis, 1998). We investigated whether action representations facilitate object recognition. Subjects were sequentially presented with two pictures of artifactual manipulable objects (e.g., tools, musical instruments) and were required to name the objects. In the congruent condition both objects afford similar actions, while they differ in the incongruent condition. Stimulus pairs in both conditions were matched for baseline naming accuracy, word frequency, word length, as well as visual and semantic similarity. Naming accuracy was significantly higher in the congruent condition. This action priming effect substantiates our hypothesis that action representations play an important role in the recognition of artifactual manipulable objects. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff Abstract Talk Biologische Kybernetik Max-Planck-Gesellschaft Mainz, Germany 48. Tagung Experimentell Arbeitender Psychologen (TeaP 2006) helbigHBHelbig grafmMGraf MKiefer conference 2452 Experimentelle Psychologie 2004 4 46 Um ein visuell wahrgenommenes Objekt greifen zu können, muss eine Umrechnung von Augenkoordinaten in Handkoordinaten erfolgen. Diese Koordinatentransformationen sind im Parietalkortex durch gain Modulation der neuronalen Antworten implementiert. Während das Modell der zwei getrennten visuellen Ströme postuliert, dass Handlungssteuerung und Objekterkennung auf unterschiedlichen Verarbeitungsprinzipien beruhen, wird hier eine integrative Sichtweise vorgestellt, nach der das Erkennen räumlich transformierter Objekte ebenfalls auf Koordinatentransformationen basiert: Durch Transformation des perzeptuellen Koordinatensystems werden Unterschiede in der räumlichen Ausrichtung zwischen visuellem Input und den Gedächtnisrepräsentationen kompensiert. Diese Konzeption wird von zwei unterschiedlichen Forschungsrichtungen gestützt. Zum einen wurde demonstriert, dass auch Objekterkennen durch gain Modulation modelliert werden kann, und dass gain Modulations-Prozesse tatsächlich im ventralen Pfad auftreten. Zum andern zeigten psychophysische Experimente, dass Objekte besser erkannt werden, wenn zuvor ein anderes Objekt in der gleichen Orientierung dargeboten wurde. Dieser Kongruenzeffekt konnte für bekannte Objekte nachgewiesen werden, selbst wenn diese visuell unähnlich sind, oder wenn die Objekte unterschiedliche Hauptachsen aufweisen. http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de http://www.kyb.tuebingen.mpg.de Department Bülthoff http://www.allpsych.uni-giessen.de/teap/programm_talks.php#ueberblick Biologische Kybernetik Max-Planck-Gesellschaft Giessen, Germany 46. Tagung Experimentell Arbeitender Psychologen (TeaP 2004) grafmMGraf