Project Leaders

 Dr. Isabelle Bülthoff
Phone: +49 7071 601-611
Fax: +49 7071 601-616

RecCat Overview Poster

Five most recent publications

Chuang LL, Gehring S, Kay J and Schmidt A: Ambient Notification Environments, Dagstuhl Seminar 17161, -, Leibniz-Zentrum für Informatik, Schloss Dagstuhl, Germany, (April-2017).in press
-, Series: Dagstuhl Reports
Chuang LL (November-5-2015) Invited Lecture: Beyond Steering in Human-Centered Closed-Loop Control, Institute for Neural Computation: INC Chalk Talk Series, San Diego, CA, USA.
Stangl M, Meilinger T, Pape A-A, Schultz J, Bülthoff HH and Wolbers T (October-19-2015): Triggers of entorhinal grid cell and hippocampal place cell remapping in humans, 45th Annual Meeting of the Society for Neuroscience (Neuroscience 2015), Chicago, IL, USA.
Fademrecht L, Bülthoff I, Barraclough NE and de la Rosa S (October-18-2015): The spatial extent of action sensitive perceptual channels decrease with visual eccentricity, 45th Annual Meeting of the Society for Neuroscience (Neuroscience 2015), Chicago, IL, USA.
Scheer M, Bülthoff HH and Chuang LL (October-2015) On the influence of steering on the orienting response In: Trends in Neuroergonomics, , 11. Berliner Werkstatt Mensch-Maschine-Systeme, Universitätsverlag der TU Berlin, Berlin, Germany, 24.

Export as:
BibTeX, XML, Pubman, Edoc, RTF

All RecCat publications

For all publications by RecCat members, click here


Recognition and Categorization

This office scene illustrates many of the problems encountered when we recognize and categorize objects. For example, we recognize all chairs easily, although they are seen from different orientations and although some of them are partially occluded.
We can easily recognize and categorize objects at different levels depending on task requirements. An animal can be recognized as belonging to a category such as “a dog” (categorization) or as “my dog Bashi” (identification). Among all categories of objects, faces constitute a very special class because of their social importance and their high intra-group similarity. Presently, the RECCAT group mainly focuses on the perception of faces.
The long-standing goal of our group is to unravel the mechanisms underlying recognition and categorization of faces and objects. Our research focuses on two main aspects:
(1) To investigate recognition and categorization in more naturalistic scenarios. we use a mix of modern and traditional methods. In many tests, faces and human figure stimuli are moving and/or shown in their natural size. In other experiments, we test face recognition more naturalistically by allowing observers to actively move around in virtual environments.
(2) To investigate perception in other populations, such as those with different cultural backgrounds (e.g. Koreans) or lacking specific skills (e.g. prosopagnosics), and more generally to investigate the role of expertise and cultural background in face recognition.

Main research areas

The role of dynamic information for object and face recognition:
Emotional expressions in general are rather easily recognized in static displays, but they represent only 10 percent of all human facial expressions. In contrast, we have shown that dynamic information conveyed in conversational expressions (eg., “I don’t understand” ) is essential for correct interpretations of those expressions [Kaulard, de la Rosa, Wallraven]. Another aspect of facial movements is how they help us to recognize the identity of a person. Facial movements of several persons were recorded and these movements were retargeted to avatar faces to investigate the role of motion idiosyncrasy for the recognition of face identity [Dobs]. Using functional brain imaging, we found that the increased response to moving compared to static faces is the result of both the fluid facial motion itself as well as the increase in static information [Schultz]. In the domain of dynamic objects, we could show that brain regions showed neural activity varying in parallel with the degree to which a single moving dot was perceived as inanimate or animate, suggesting that biological motion processing and detection of animacy are implemented in different brain areas [Schultz].
The role of the observer's actions on perception:
View generalization for faces was tested in a virtual setup that allowed participants to move freely and collect actively numerous views of sitting and standing avatars that they encountered. Despite that, view generalization along the pitch axis did not occur for avatar faces, indicating that view dependency is not the consequence of passive viewing [I. Bülthoff]. Furthermore, data collected from active observers manipulating virtual objects revealed that observers rely on non-accidental properties of the objects (i.e., symmetry and elongation) to select views that they used for subsequent recognition [Chuang, Wallraven].
Cross-modal perception:
Using haptic objects (plastic 3D face masks and sea shells) we have shown that training in one sensory modality can transfer to another modality, resulting in improved discrimination performance [Dopjans, GaißertWallraven]. This transfer indicates that vision and touch share similar representations. Furthermore, we have shown for the first time that categorical perception also occurs in haptic shape perception. Participants performed a classification and discrimination task using a continuum of morphed plastic objects. A brief training altered haptic representations of shape resulting in categorical perception.
Populations with atypical perception:
In accordance with previous studies, we have shown that Korean observers fixate faces differently than their German counterparts, but interestingly, despite those differences in gaze movements, we found no effect of own-face expertise in terms of accuracy performance when the tasks did not involve memorizations of facial identity [I. Bülthoff]. We have started to investigate a large population of congenital Prosopagnosics. Preliminary results indicate that they are less sensitive to configural than to featural changes in faces that were modified parametrically. The use of a parametric stimulus space will allow a sensitive and precise quantitative description of their deficits [Esins]. Using parametric motion stimuli varying in complexity from simple translational motion to interacting shapes, we found that people with Autism Spectrum Disorder show deficits in assessing social interactions between moving objects [Schultz]. Faces were also used in a project to give a definite answer to the controversy about whether faces are perceived categorically in terms of their sex or not. Using our face database and the morphable model to manipulate the face stimuli to obtain very similar faces in terms of perceived masculinity and feminity allowed us to show that sex is not perceived categorically [Armann].

Methods and Stimuli

Example stimuli used in our group. A: static image from a conversational facial expression (“don’t know”). B: Asian version of a Caucasian face. C: avatar face animated by the smile of a real person. D: 3D-printout of a face of our 3D face database. E: artificial symmetric object. F: Trajectory of a simulated fly. G: 3D-printout of a parametrically-controlled shell object. Stimuli B, C, D, E, F and G can be parametrically modified.
Dynamic faces are recorded or created using the MPI VideoLab, as well as our in-house facial animation software developed by the Cognitive Engineering group [Curio].
Active behavior research is using both body-tracking technology (VICON MX) and immersive head-mounted visual displays.
Real volumetric objects for haptic and vision research are produced with 3D printing technology.
Virtual reality setups are used to present faces and objects embedded in scenes and allow active manipulations of objects.
Furthermore, classical psychophysical techniques, functional brain imaging  [Siemens 3T at Magnetic Resonance Center] and eye tracking methods are used in the projects of the group.


The Cognitive Neuroimaging group [working with Armann, Schultz], the Physiology of Cognitive Processes department [Schultz], the Cognitive Engineering group [DobsChuang] and the PAVE group [I. Bülthoff] were major partners in the development and the execution of many projects. Other collaborators reside outside of the institute: e.g., Christian Wallraven (Korea), Ian M Thornton (UK), Karin S Pilz (Switzerland), Nicole David (Germany), Quoc Vuong (UK). Several projects of the group were part of the EU-funded project POETICON.


For a full list of publications by RecCat members, click here

Selected publications: Articles 2009-present

Bieg H-J, Bresciani J-P, Bülthoff HH and Chuang LL (September-2012) Looking for Discriminating Is Different from Looking for Looking's Sake PLoS ONE 7(9) 1-9.
Gaissert N, Waterkamp S, Fleming RW and Bülthoff I (August-2012) Haptic Categorical Perception of Shape PLoS One 7(8) 1-7.
Bülthoff I (July-2012) Review: L'empreinte Des Sens Perception 41(7) 881-882.
Armann R and Bülthoff I (June-2012) Male and female faces are only perceived categorically when linked to familiar identities – And when in doubt, he is a male Vision Research 63 69–80.
Chuang LL, Vuong QC and Bülthoff HH (May-2012) Learned non-rigid object motion is a view-invariant cue to recognizing novel objects Frontiers in Computational Neuroscience 6(26) 1-8.
Helbig HB, Ernst MO, Ricciardi E, Pietrini P, Thielscher A, Mayer KM, Schultz J and Noppeney U (April-2012) The neural mechanisms of reliability weighted integration of shape information from vision and touch NeuroImage 60(2) 1063–1072.
Kaulard K, Cunningham DW, Bülthoff HH and Wallraven C (March-2012) The MPI Facial Expression Database: A Validated Database of Emotional and Conversational Facial Expressions PLoS One 7(3) 1-18.
Gaissert N and Wallraven C (January-2012) Categorizing natural objects: a comparison of the visual and the haptic modalities Experimental Brain Research 216(1) 123-134.
Dopjans L, Bülthoff HH and Wallraven C (January-2012) Serial exploration of faces: Comparing vision and touch Journal of Vision 12(1:6) 1-14.
Armann R, Jeffery L, Calder AJ and Rhodes G (November-2011) Race-specific norms for coding face identity and a functional role for norms Journal of Vision 11(13:9) 1-14.
Bülthoff HH and Chuang LL (September-2011) Seeing: The Computational Approach to Biological Vision. Second Edition. By John P. Frisby and James V. Stone. Cambridge (Massachusetts): MIT Press Quarterly Review of Biology 86(3) 227.
Gaissert N, Bülthoff HH and Wallraven C (September-2011) Similarity and categorization: From vision to touch Acta Psychologica 138(1) 219-230.
1, 2

Export as:
BibTeX, XML, Pubman, Edoc, RTF
Last updated: Monday, 04.05.2015