Neuroscience and Neurobiology Commons^™

Derivatives And Inverse Of A Linear-Nonlinear Multi-Layer Spatial Vision Model, Borja Galan, Marina Martinez-Garcia, Praveen Cyriac, Thomas Batard, Marcelo Bertalmio, Jesus Malo

MODVIS Workshop

Analyzing the mathematical properties of perceptually meaningful linear-nonlinear transforms is interesting because this computation is at the core of many vision models. Here we make such analysis in detail using a specific model [Malo & Simoncelli, SPIE Human Vision Electr. Imag. 2015] which is illustrative because it consists of a cascade of standard linear-nonlinear modules. The interest of the analytic results and the numerical methods involved transcend the particular model because of the ubiquity of the linear-nonlinear structure.

Here we extend [Malo&Simoncelli 15] by considering 4 layers: (1) linear spectral integration and nonlinear brightness response, (2) definition of local contrast …

Towards A Functional Explanation Of The Connectivity Lgn - V1, Marina Martinez-Garcia, Borja Galan, Luis M. Martinez, Jesus Malo

MODVIS Workshop

The principles behind the connectivity between LGN and V1 are not well understood. Models have to explain two basic experimental trends: (i) the combination of thalamic responses is local and it gives rise to a variety of oriented Gabor-like receptive felds in V1 [1], and (ii) these filters are spatially organized in orientation maps [2]. Competing explanations of orientation maps use purely geometrical arguments such as optimal wiring or packing from LGN [3-5], but they make no explicit reference to visual function. On the other hand, explanations based on func- tional arguments such as maximum information transference (infomax) [6,7] usually …

Towards A Unified Model Of Classical And Extra-Classical Receptive Fields, David A. Mély, Thomas Serre

MODVIS Workshop

One of the major goals in neuroscience is to understand how the cortex processes information. A substantial effort has thus gone into mapping classical receptive fields (cRF) across areas of the visual cortex and characterizing input-output relationships through linear-nonlinear response functions. Recently, there has been a lot of interest in mapping the extra-classical receptive field (extra-cRF) as well, by using contextual stimuli. The extra-cRF is a region outside the cRF that modulates a cell’s response but that is incapable of driving it on its own. However, existing models typically focus on one particular visual modality (form, motion, disparity or color), …

How Deep Is The Feature Analysis Underlying Rapid Visual Categorization?, Sven Eberhardt, Jonah Cader, Thomas Serre

MODVIS Workshop

Rapid categorization paradigms have a long history in experimental psychology: Characterized by short presentation times and fast behavioral responses, these tasks highlight both the speed and ease with which our visual system processes natural object categories. Previous studies have shown that feed-forward hierarchical models of the visual cortex provide a good fit to human visual decisions. At the same time, recent work has demonstrated significant gains in object recognition accuracy with increasingly deep hierarchical architectures: From AlexNet to VGG to Microsoft CNTK – the field of computer vision has championed both depth and accuracy. But it is unclear how well …

Spatial Synaptic Growth And Removal For Learning Individual Receptive Field Structures, Michael Teichmann, Fred H. Hamker

MODVIS Workshop

One challenge in creating neural models of the visual system is the appropriate definition of the connectivity. The modeler constrains the results with its definition. Unfortunately, there is often just insufficient information about connection sizes available, e.g. for deeper layer or different neuron types like interneurons. Hence, a mechanism refining the connection structure based on the learnings would be appreciated.

Such mechanism can be found in the human brain by structural plasticity. That is, the formation and removal of synapses. For our model, we exploit that synaptic connections are likely to be formed in the proximity of other synapses and …

A Geometric Approach To Sparse Coding Yields Insight Into Nonlinear Responses, Kedarnath Vilankar, James Golden, David Field

MODVIS Workshop

In artificial and biological networks, it is a common accepted practice to describe a neurons (biological or artificial) response properties by a two-dimensional feature map (receptive field). However, real neurons have nonlinear response properties which are not represented by their receptive fields. The efficient coding mechanisms such as sparse coding network or ICA, learn the response properties of V1 neurons from natural images using neural networks. These networks learn the receptive fields which are similar to the receptive fields of V1 neurons. These networks also produces some of the nonlinearities (such as end-stopping and non-classical surround effect), which are exhibited …

Using Deep Features To Predict Where People Look, Matthias Kümmerer, Matthias Bethge

MODVIS Workshop

When free-viewing scenes, the first few fixations of human observers are driven in part by bottom-up attention. We seek to characterize this process by extracting all information from images that can be used to predict fixation densities (Kuemmerer et al, PNAS, 2015). If we ignore time and observer identity, the average amount of information is slightly larger than 2 bits per image for the MIT 1003 dataset. The minimum amount of information is 0.3 bits and the maximum 5.2 bits. Before the rise of deep neural networks the best models were able to capture 1/3 of this information on average. …

Modelling Short-Latency Disparity-Vergence Eye Movements Under Dichoptic Unbalanced Stimulation, Agostino Gibaldi, Guido Maiello, Peter J. Bex, Silvio P. Sabatini

MODVIS Workshop

Vergence eye movements align the optical axes of our two eyes onto an object of interest, thus facilitating the binocular summation of the images projected onto the left and the right retinae into a single percept. Both the computational substrate and the functional behaviour of binocular vergence eye movements have been the topic of in depth investigation. Here, we attempt to bring together what is known about computation and function of vergence mechanism. To this aim, we evaluated of a biologically inspired model of horizontal and vertical vergence control, based on a network of V1 simple and complex cells. The …

Learning Object Representations For Modeling Attention In Real World Scenes, Alex Schwarz, Frederik Beuth, Fred H. Hamker

MODVIS Workshop

Models of visual attention have been rarely used in real world tasks as they have been typically developed for psychophysical setups using simple stimuli. Thus, the question remains how objects must be represented to allow such models an operation in real world scenarios. We have previously presented an attention model capable of operating on real-world scenes (Beuth, F., and Hamker, F. H. 2015, NCNC, which is a successor of Hamker, F. H., 2005, Cerebral Cortex), and show here how its object representations have been learned. We have used a learning rule based on temporal continuity (Földiák, P., 1991, Neural Computation) …

Characterizing Receptive Field Selectivity In Area V2, Corey M. Ziemba, Robbe Lt Goris, J Anthony Movshon, Eero P. Simoncelli

MODVIS Workshop

The computations performed by neurons in area V1 are reasonably well understood, but computation in subsequent areas such as V2 have been more difficult to characterize. When stimulated with visual stimuli traditionally used to investigate V1, such as sinusoidal gratings, V2 neurons exhibit similar selectivity (but with larger receptive fields, and weaker responses) relative to V1 neurons. However, we find that V2 responses to synthetic stimuli designed to produce naturalistic patterns of joint activity in a model V1 population are more vigorous than responses to control stimuli that lacked this naturalistic structure (Freeman, et. al. 2013). Armed with this signature …

‘Edge’ Integration Explains Contrast And Assimilation In A Gradient Lightness Illusion, Michael E. Rudd

MODVIS Workshop

In the ‘phantom’ illusion (Galmonte, Soranzo, Rudd, & Agostini, submitted), either an incremental or a decremental target, when surrounded by a luminance gradient, can to be made to appear as an increment or a decrement, depending on the gradient width. For wide gradients, incremental targets appear as increments and decremental targets appear as decrements. For narrow gradients, the reverse is true. Here, I model these phenomena with a two-stage neural lightness theory (Rudd, 2013, 2014) in which local steps in log luminance are first encoded by oriented spatial filters operating on a log-transformed version of the image; then the filter …

Time-Lapse Statistics Of Cone Signals From Natural Scenes, David H. Foster, Kinjiro Amano, Sérgio M C Nascimento

MODVIS Workshop

In the natural world, the spectrum and geometry of the illumination from the sun and sky vary over the day. These geometric changes make it especially difficult to extract from the reflected light invariant signals for surface color perception. The aim of this study was to test the utility of certain combinations of retinal cone excitations, in particular, spatial cone-excitation ratios, known to be approximately invariant under non-geometric changes in illumination. Hyperspectral radiance images were acquired at roughly hourly intervals in four outdoor scenes in the Minho region of Portugal. Spatial resolution of the camera was 1344×1024 pixels. Peak-transmission wavelength …

The Bounded Log-Odds Model Of Frequency And Probability Distortion, Hang Zhang, Laurence T. Maloney

MODVIS Workshop

No abstract provided.

Putting Saliency In Its Place, John K. Tsotsos

MODVIS Workshop

The role of attention and the place within the visual processing stream where the concept of saliency has been situated is critically examined by considering the experimental evidence and performing tests that link experiment to computation.

Towards A Unified Computational Model Of Contextual Interactions Across Visual Modalities, David A. Mély, Thomas Serre

MODVIS Workshop

The perception of a stimulus is largely determined by its surrounding. Examples abound from color (Land and McCann, 1971), disparity (Westheimer, 1986) and motion induction (Anstis and Casco, 2006) to orientation tilt effects (O’Toole and Wenderoth, 1976). Some of these phenomena have been studied individually using monkey neurophysiology techniques. In these experiments, a center stimulus is typically used to probe a cell’s classical “center” receptive field (cRF), whose activity is then modulated by an annular “surround” (extra-cRF) stimulus. While this center-surround integration (CSI) has been well characterized, a theoretical framework which unifies these different phenomena across visual modalities is lacking. …

A Conceptual Framework Of Computations In Mid-Level Vision, Jonas Kubilius, Johan Wagemans, Hans P. Op De Beeck

MODVIS Workshop

The goal of visual processing is to extract information necessary for a variety of tasks, such as grasping objects, navigating in scenes, and recognizing them. While ultimately these tasks might be carried out by separate processing pathways, they nonetheless share a common root in the early and intermediate visual areas. What representations should these areas develop in order to facilitate all of these higher-level tasks? Several distinct ideas have received empirical support in the literature so far: (i) boundary feature detection, such as edge, corner, and curved segment extraction; (ii) second-order feature detection, such as the difference in orientation or …

Binocular 3d Motion Perception As Bayesian Inference, Martin Lages, Suzanne Heron

MODVIS Workshop

The human visual system encodes monocular motion and binocular disparity input before it is integrated into a single 3D percept. Here we propose a geometric-statistical model of human 3D motion perception that solves the aperture problem in 3D by assuming that (i) velocity constraints arise from inverse projection of local 2D velocity constraints in a binocular viewing geometry, (ii) noise from monocular motion and binocular disparity processing is independent, and (iii) slower motions are more likely to occur than faster ones. In two experiments we found that instantiation of this Bayesian model can explain perceived 3D line motion direction under …

A Binocular Model For Motion Integration In Mt Neurons, Pamela M. Baker, Wyeth Bair

MODVIS Workshop

Processing of visual motion by neurons in MT has long been an active area of study, however circuit models detailing the computations underlying binocular integration of motion signals remains elusive. Such models are important for studying the visual perception of motion in depth (MID), which involves both frontoparallel (FP) visual motion and binocular signal integration. Recent studies (Czuba et al. 2014, Sanada and DeAngelis 2014) have shown that many MT neurons are MID sensitive, contrary to the prevailing view (Maunsell and van Essen, 1983). These novel data are ideal for constraining models of binocular motion integration in MT. We have …

A Critical Evaluation Of Computational Mechanisms Of Binocular Disparity Processing, Junkyung Kim, David A. Mély, Thomas Serre

MODVIS Workshop

The past decades of research in visual neuroscience have generated a large and disparate body of literature on the computation of binocular disparity in the primary visual cortex. Models have been proposed to describe specific phenomena, yet we lack a theoretical framework which is grounded in neurophysiology and also explains the effectiveness of disparity computation. Here, we examine neural circuits that are thought to play an important role in the computation of binocular disparity. Starting with the binocular energy model (Ohzawa et al. 1990), we consider plausible extensions which include suppressive mechanisms from units tuned to different phase disparities (Tanabe …

Modeling Shape Representation In Area V4, Wyeth Bair, Dina Popovkina, Abhishek De, Anitha Pasupathy

MODVIS Workshop

Our model builds on a convolutional-style neural network with hierarchical stages representing processing steps in the ventral visual pathway. It was designed to capture the translation-invariance and shape-selectivity of neurons in area V4. The model uses biologically plausible linear filters at the front end, normalization and sigmoidal nonlinear activation functions. The max() function is used to generate translation invariance.

A Recurrent Multilayer Model With Hebbian Learning And Intrinsic Plasticity Leads To Invariant Object Recognition And Biologically Plausible Receptive Fields, Michael Teichmann, Fred H. Hamker

MODVIS Workshop

Much effort has been spent to develop biologically plausible models for different aspects of the processing in the visual cortex. However, most of these models are not investigated with respect to the functionality of the neural code for the purpose of object recognition comparable to the framework of deep learning in the machine learning community.
We developed a model of V1 and V2 based on anatomical evidence of the layered architecture, using excitatory and inhibitory neurons where the connectivity to each neuron is learned in parallel. We address learning by three different mechanisms of plasticity: intrinsic plasticity, Hebbian learning with …

A Computational Model To Account For Dynamics Of Spatial Updating Of Remembered Visual Targets Across Slow And Rapid Eye Movements, Yalda Mohsenzadeh, J. Douglas Crawford

MODVIS Workshop

Despite the ever-changing visual scene on the retina between eye movements, our perception of the visual world is constant and unified. It is generally believed that this space constancy is due to the brain’s ability of spatial updating. Although many efforts have been made to discover the mechanism underlying spatial updating across eye movements, still there are many unanswered questions about the neuronal mechanism of this phenomenon.

We developed a state space model for updating gaze-centered spatial information. To explore spatial updating, we considered two kinds of eye movements, saccade and smooth pursuit. The inputs to our proposed model are: …

A Model Of Repetitive Microsaccades, Coupled With Pre-Microsaccadic Changes In Vision, Is Sufficient To Account For Both Attentional Capture And Inhibition Of Return In Posner Cueing, Ziad M. Hafed, Xiaoguang Tian

MODVIS Workshop

When a cue is presented at a location, orienting efficacy towards that location is improved relative to other locations (“attentional capture”), but only briefly; a mere few hundred milliseconds later, orienting incurs large costs. These costs have been classically termed “inhibition of return” (IOR), alluding to voluntary, cognitive strategies avoiding perseverance at one location. However, despite this popular hypothesis, the origins of both attentional capture and IOR remain elusive. Here we show that both of these phenomena can be accounted for by a single concept of oculomotor rhythmicity, and one that involves the entire gamut of saccadic activity including microsaccades. …

Object Recognition And Visual Search With A Physiologically Grounded Model Of Visual Attention, Frederik Beuth, Fred H. Hamker

MODVIS Workshop

Visual attention models can explain a rich set of physiological data (Reynolds & Heeger, 2009, Neuron), but can rarely link these findings to real-world tasks. Here, we would like to narrow this gap with a novel, physiologically grounded model of visual attention by demonstrating its objects recognition abilities in noisy scenes.

To base the model on physiological data, we used a recently developed microcircuit model of visual attention (Beuth & Hamker, in revision, Vision Res) which explains a large set of attention experiments, e.g. biased competition, modulation of contrast response functions, tuning curves, and surround suppression. Objects are represented by …

Modeling Visual Features To Recognize Biological Motion: A Developmental Approach, Giulio Sandini, Nicoletta Noceti, Alessia Vignolo, Alessandra Sciutti, Francesco Rea, Alessandro Verri, Francesca Odone

MODVIS Workshop

In this work we deal with the problem of designing and developing computational vision models – comparable to the early stages of the human development – using coarse low-level information.

More specifically, we consider a binary classification setting to characterize biological movements with respect to non-biological dynamic events. To this purpose, our model builds on top of the optical flow estimation, and abstract the representation to simulate the limited amount of visual information available at birth. We take inspiration from known biological motion regularities explained by the Two-Thirds Power Law, and design a motion representation that includes different low-level features, …