The utility of shape attributes in deciphering movements of non-rigid objects

Abstract

Most moving objects in the world are non-rigid, changing shape as they move. To disentangle shape changes from movements, computational models either fit shapes to combinations of basis shapes or motion trajectories to combinations of oscillations but are biologically unfeasible in their input requirements. Recent neural models parse shapes into stored examples, which are unlikely to exist for general shapes. We propose that extracting shape attributes, e.g., symmetry, facilitates veridical perception of non-rigid motion. In a new method, identical dots were moved in and out along invisible spokes, to simulate the rotation of dynamically and randomly distorting shapes. Discrimination of rotation direction measured as a function of non-rigidity was 90% as efficient as the optimal Bayesian rotation decoder and ruled out models based on combining the strongest local motions. Remarkably, for non-rigid symmetric shapes, observers outperformed the Bayesian model when perceived rotation could correspond only to rotation of global symmetry, i.e., when tracking of shape contours or local features was uninformative. That extracted symmetry can drive perceived motion suggests that shape attributes may provide links across the dorsal–ventral separation between motion and shape processing. Consequently, the perception of non-rigid object motion could be based on representations that highlight global shape attributes.

Efficiency of extracting stereo-driven object motions

“…The efficiency for 3-D shapes is lower than for 2-D deforming objects made of orthogonal local motions Cohen et al (2010). For 2-D objects, observers were 90 % as efficient as an optimal Bayesian decoder and even out-performed the decoder when the shapes were symmetric.…”

“…In such cases, it is possible that information from the fovea region is weighted more than information in periphery due to a decline in stereo-acuity with eccentricity (Cumming & DeAngelis, 2001; Parker, 2007; Wardle, Bex, Cass, & Alais, 2012). Further, 2D shapes formed by visible patches can be used to extract motion as well and Cohen et al (2010) have shown that humans are not only extremely efficient at using 2D shapes but can also use abstract properties such as symmetry to extract object motion.…”

“…If a moving contour is visible, inferences can be made about shape (Cipolla & Giblin, 2000) and motion (Caplovitz & Tse, 2007a, 2007b; Rokers, Yuille, & Liu, 2006). If the object is partially occluded so that the contour is sparsely sampled, shape properties can help to infer motion direction for rigid (Lorenceau & Alais, 2001; Shiffrar & Pavel, 1991) and non-rigid (Cohen et al, 2010) objects, and motion can reveal shapes through pattern integration (Nishida, 2004). If the contour is completely occluded, visible patterns of velocities can be used to perceive 3-D shape for rigid (Koenderink & van Doorn, 1975, 1991) and non-rigid (Akhter, Sheikh, Khan, & Kanade, 2008; Bregler, Hertzmann, & Biermann, 2000; Jain & Zaidi, 2011) objects.…”

“…When the object is partially occluded, the task becomes more difficult because it is not possible to use 2-D contour correlations (Cohen, Jain, & Zaidi, 2010). That leaves dynamic depth matching across the un-occluded views as the main possibility.…”

Mental geometry of 3D size and shape perception

“…In the dynamic case, this could be related to the possibility that articulated objects are more common in the world then are objects that increase or decrease in length over a short period. The same appeal to natural statistics could be invoked to explain the difficulty in perceiving expansion or contraction in depth of solid objects (Johansson, 1964;Jansson & Johansson, 1973;Jain & Zaidi, 2011), while other deformations are easy to discern, even for rotating and flowing shapes (Cohen, Jain & Zaidi, 2010;Fantoni, Caudek & Domini, 2014;Bates et al, 2019). However, the slant illusion is just as compelling in the static case, so biases in perceiving depth versus extent from retinal images may also be in play (Jain & Zaidi, 2013;Kim & Burge 2018).…”