Learning features of intermediate complexity for the recognition of biological motion.pdf

W. Duch et al. (Eds.): ICANN 2005, LNCS 3696, pp. 241 – 246, 2005. ? Springer-Verlag Berlin Heidelberg 2005 Learning Features of Intermediate Complexity for the Recognition of Biological Motion Rodrigo Sigala1,2, Thomas Serre2, Tomaso Poggio2, and Martin Giese1 1 Laboratory for Action Representation and Learning (ARL), Dept. of Cognitive Neurology, University Clinic Tübingen, Schaffhausenstr. 113, D-72072 Tübingen, Germany Rodrigo.Sigala@tuebingen.mpg.de, Martin.Giese@uni-tuebingen.de 2 McGovern Institute for Brain Research, Brain and Cognitive Sciences, Massachusetts Institute of Technology, Bldg.E25-201, 45 Carleton St., Cambridge, MA 02142, USA serre@mit.edu, poggio@ai.mit.edu Abstract. Humans can recognize biological motion from strongly impoverished stimuli, like point-light displays. Although the neural mechanism underlying this robust perceptual process have not yet been clarified, one possible explana- tion is that the visual system extracts specific motion features that are suitable for the robust recognition of both normal and degraded stimuli. We present a neural model for biological motion recognition that learns robust mid-level mo- tion features in an unsupervised way using a neurally plausible memory-trace learning rule. Optimal mid-level features were learnt from image motion se- quences containing a walker with, or without background motion clutter. After learning of the motion features, the detection performance of the model sub- stantially increases, in particular in presence of clutter. The learned mid-level motion features are characterized by horizontal opponent motion, where this feature type arises more frequently for the training stimuli without motion clut- ter. The learned features are consistent with recent psychophysical data that in- dicates that opponent motion might be critical for the detection of point light walkers. 1 Introduction Humans can recognize biological motion (e.g. human actions like walking an