A promising approach to image processing based on neuromorphic decoding in the Marr's paradigm

Abstract

In computer vision tasks, detecting object boundaries, as well as their textures, is one of the key ones. Despite the fact that significant progress has already been made in object recognition tasks, existing processing models and algorithms are significantly inferior to the capabilities of the visual system. In previous works, several methods have been proposed to determine the contours of objects, allowing not only to highlight the boundaries of each object in the image, but also to significantly eliminate distortions associated with the blurring of fuzzy images, reduce the noise component of the image, and restore the indistinctly defined border. This article discusses current issues in computer vision using the example of a previously developed bioinspired image processing algorithm.