A model of neural mechanisms in monocular transparent motion perception

Canonical circuit computations for computer vision

Advanced computer vision mechanisms have been inspired by neuroscientific findings. However, with the focus on improving benchmark achievements, technical solutions have been shaped by application and engineering constraints. This includes the training of neural networks which led to the development of feature detectors optimally suited to the application domain. However, the limitations of such approaches motivate the need to identify computational principles, or motifs, in biological vision that can enable further foundational advances in machine vision. We propose to utilize structural and functional principles of neural systems that have been largely overlooked. They potentially provide new inspirations for computer vision mechanisms and models. Recurrent feedforward, lateral, and feedback interactions characterize general principles underlying processing in mammals. We derive a formal specification of core computational motifs that utilize these principles. These are combined to define model mechanisms for visual shape and motion processing. We demonstrate how such a framework can be adopted to run on neuromorphic brain-inspired hardware platforms and can be extended to automatically adapt to environment statistics. We argue that the identified principles and their formalization inspires sophisticated computational mechanisms with improved explanatory scope. These and other elaborated, biologically inspired models can be employed to design computer vision solutions for different tasks and they can be used to advance neural network architectures of learning.

Modeling Physiological Sources of Heading Bias from Optic Flow

Human heading perception from optic flow is accurate for directions close to the straight-ahead and systematic biases emerge in the periphery (Cuturi and Macneilage, 2013; Sun et al., 2020). In pursuit of the underlying neural mechanisms, primate brain dorsal medial superior temporal (MSTd) area has been a focus because of its causal link with heading perception (Gu et al., 2012). Computational models generally explain heading sensitivity in individual MSTd neurons as a feedforward integration of motion signals from medial temporal (MT) area that resemble full-field optic flow patterns consistent with the preferred heading direction (Britten, 2008; Mineault et al., 2012). In the present simulation study, we quantified within the structure of this feedforward model how physiological properties of MT and MSTd shape heading signals. We found that known physiological tuning characteristics generally supported the accuracy of heading estimation, but not always. A weak-to-moderate overrepresentation of peripheral headings in MSTd garnered the highest accuracy and precision out of the models that we tested. The model also performed well when noise corrupted high proportions of the optic flow vectors. Such a peripheral MSTd model performed well when units possessed a range of receptive field (RF) sizes and were strongly direction tuned. Physiological biases in MT direction tuning toward the radial direction also supported heading estimation, but the tendency for MT preferred speed and RF size to scale with eccentricity did not. Our findings help elucidate the extent to which different physiological tuning properties influence the accuracy and precision of neural heading signals.

The Role of Bottom-Up and Top-Down Cortical Interactions in Adaptation to Natural Scene Statistics

Adaptation is a mechanism by which cortical neurons adjust their responses according to recently viewed stimuli. Visual information is processed in a circuit formed by feedforward (FF) and feedback (FB) synaptic connections of neurons in different cortical layers. Here, the functional role of FF-FB streams and their synaptic dynamics in adaptation to natural stimuli is assessed in psychophysics and neural model. We propose a cortical model which predicts psychophysically observed motion adaptation aftereffects (MAE) after exposure to geometrically distorted natural image sequences. The model comprises direction selective neurons in V1 and MT connected by recurrent FF and FB dynamic synapses. Psychophysically plausible model MAEs were obtained from synaptic changes within neurons tuned to salient direction signals of the broadband natural input. It is conceived that, motion disambiguation by FF-FB interactions is critical to encode this salient information. Moreover, only FF-FB dynamic synapses operating at distinct rates predicted psychophysical MAEs at different adaptation time-scales which could not be accounted for by single rate dynamic synapses in either of the streams. Recurrent FF-FB pathways thereby play a role during adaptation in a natural environment, specifically in inducing multilevel cortical plasticity to salient information and in mediating adaptation at different time-scales.

Cross-Modal Cue Effects in Motion Processing

The everyday environment brings to our sensory systems competing inputs from different modalities. The ability to filter these multisensory inputs in order to identify and efficiently utilize useful spatial cues is necessary to detect and process the relevant information. In the present study, we investigate how feature-based attention affects the detection of motion across sensory modalities. We were interested to determine how subjects use intramodal, cross-modal auditory, and combined audiovisual motion cues to attend to specific visual motion signals. The results showed that in most cases, both the visual and the auditory cues enhance feature-based orienting to a transparent visual motion pattern presented among distractor motion patterns. Whereas previous studies have shown cross-modal effects of spatial attention, our results demonstrate a spread of cross-modal feature-based attention cues, which have been matched for the detection threshold of the visual target. These effects were very robust in comparisons of the effects of valid vs. invalid cues, as well as in comparisons between cued and uncued valid trials. The effect of intramodal visual, cross-modal auditory, and bimodal cues also increased as a function of motion-cue salience. Our results suggest that orienting to visual motion patterns among distracters can be facilitated not only by intramodal priors, but also by feature-based cross-modal information from the auditory system.

On event-based optical flow detection

Event-based sensing, i.e., the asynchronous detection of luminance changes, promises low-energy, high dynamic range, and sparse sensing. This stands in contrast to whole image frame-wise acquisition by standard cameras. Here, we systematically investigate the implications of event-based sensing in the context of visual motion, or flow, estimation. Starting from a common theoretical foundation, we discuss different principal approaches for optical flow detection ranging from gradient-based methods over plane-fitting to filter based methods and identify strengths and weaknesses of each class. Gradient-based methods for local motion integration are shown to suffer from the sparse encoding in address-event representations (AER). Approaches exploiting the local plane like structure of the event cloud, on the other hand, are shown to be well suited. Within this class, filter based approaches are shown to define a proper detection scheme which can also deal with the problem of representing multiple motions at a single location (motion transparency). A novel biologically inspired efficient motion detector is proposed, analyzed and experimentally validated. Furthermore, a stage of surround normalization is incorporated. Together with the filtering this defines a canonical circuit for motion feature detection. The theoretical analysis shows that such an integrated circuit reduces motion ambiguity in addition to decorrelating the representation of motion related activations.

Hierarchical representation of shapes in visual cortex-from localized features to figural shape segregation

Visual structures in the environment are segmented into image regions and those combined to a representation of surfaces and prototypical objects. Such a perceptual organization is performed by complex neural mechanisms in the visual cortex of primates. Multiple mutually connected areas in the ventral cortical pathway receive visual input and extract local form features that are subsequently grouped into increasingly complex, more meaningful image elements. Such a distributed network of processing must be capable to make accessible highly articulated changes in shape boundary as well as very subtle curvature changes that contribute to the perception of an object. We propose a recurrent computational network architecture that utilizes hierarchical distributed representations of shape features to encode surface and object boundary over different scales of resolution. Our model makes use of neural mechanisms that model the processing capabilities of early and intermediate stages in visual cortex, namely areas V1-V4 and IT. We suggest that multiple specialized component representations interact by feedforward hierarchical processing that is combined with feedback signals driven by representations generated at higher stages. Based on this, global configurational as well as local information is made available to distinguish changes in the object's contour. Once the outline of a shape has been established, contextual contour configurations are used to assign border ownership directions and thus achieve segregation of figure and ground. The model, thus, proposes how separate mechanisms contribute to distributed hierarchical cortical shape representation and combine with processes of figure-ground segregation. Our model is probed with a selection of stimuli to illustrate processing results at different processing stages. We especially highlight how modulatory feedback connections contribute to the processing of visual input at various stages in the processing hierarchy.