Computational modeling and exploration of contour integration for visual saliency

Abstract

We propose a computational model of contour integration for visual saliency. The model uses biologically plausible devices to simulate how the representations of elements aligned collinearly along a contour in an image are enhanced. Our model adds such devices as a dopamine-like fast plasticity, local GABAergic inhibition and multi-scale processing of images. The fast plasticity addresses the problem of how neurons in visual cortex seem to be able to influence neurons they are not directly connected to, for instance, as observed in contour closure effect. Local GABAergic inhibition is used to control gain in the system without using global mechanisms which may be non-plausible given the limited reach of axonal arbors in visual cortex. The model is then used to explore not only its validity in real and artificial images, but to discover some of the mechanisms involved in processing of complex visual features such as junctions and end-stops as well as contours. We present evidence for the validity of our model in several phases, starting with local enhancement of only a few collinear elements. We then test our model on more complex contour integration images with a large number of Gabor elements. Sections of the model are also extracted and used to discover how the model might relate contour integration neurons to neurons that process end-stops and junctions. Finally, we present results from real world images. Results from the model suggest that it is a good current approximation of contour integration in human vision. As well, it suggests that contour integration mechanisms may be strongly related to mechanisms for detecting end-stops and junction points. Additionally, a contour integration mechanism may be involved in finding features for objects such as faces. This suggests that visual cortex may be more information efficient and that neural regions may have multiple roles.

Enhancement of Perceptually Salient Contours using a Parallel Artificial Cortical Network

“…In every image, some of the line segments form a perceptually salient contour. Similar test images have been used in many previous works (Shashua and Ullman 1988;Alter and Basri 1998;Finkel 1997, 1998;Li 1998;Choe and Miikkulainen 2004;Mundhenk and Itti 2005). The first column depicts the original image and the second column the output of the network after 10 iterations.…”

Image Segmentation Using a Sparse Coding Model of Cortical Area V1

“…For this reason, numerous neural network approaches have been proposed that are inspired by the physiology of the primary visual cortex (Ben-Shahar and Zucker, 2004;Grigorescu et al, 2003Grigorescu et al, , 2004Hansen and Neumann, 2008;Huang et al, 2009;Li, 1998;Mundhenk and Itti, 2005;Papari et al, 2007;Papari and Petkov, 2011;Petkov and Westenberg, 2003;Tang et al, 2007a,b;Ursino and La Cara, 2004;Vonikakis et al, 2006;Zeng et al, 2011a,b). A neuron in such a model has a classical receptive field (cRF), often defined using a Gabor function, that receives input from the image.…”

Edge and line oriented contour detection: State of the art

“…An important example is given by the algorithms to extract local edge features, which are not designed in terms of subsequent steps such as computing a contour saliency from local responses, or grouping edges according to Gestalt principles. In particular, much research has been made to distinguish from long chains of collinear edges and randomly scattered oriented stimuli [136,137,291]. Biological evidence suggests that the human visual system deploys similar mechanisms [139,140].…”

“…Recent work [136,137] shows that schemes similar to tensor voting can successfully model the facilitation and contour integration processes that are performed in the front-end part of the human visual system. This part of the brain is modeled as a 3D grid of neurons, which react to oriented stimuli.…”