Analyzing the Efficiency of Segment Boundary Detection Using Neural Networks

Proceedings of the 32nd International Conference on Computer Graphics and Vision

Mangarakov

2022

The article presents a new model of the MT neuron (neuron of the middle temporal region), which allows motion detecting and determining its direction and speed without the use of recurrent communication. The model is based on signal accumulation and is organized using a space-time vector that sets the weighting coefficients. The space-time vector is formed using the product of the Gaussian, which defines the spatial component, and the "Mexican hat" wavelet, which sets the time vector of the change in the receptive field. This configuration allows not only to motion detect, but also to make the model not sensitive to uniform or textural fill. The model is presented in variants for determining linear and rotational motion. Motion, in this case, is the sequential activation of several edge selection neurons located in the same direction in a certain neighborhood over time i.e. with a change of frame. To assess the motion, the models were tested on the MPI Sintel dataset. The model developed by us shows results better than Spatio-Temporal Gabor. The best accuracy of determining the direction of movement can be obtained with the size of the space-time vector (7*7, 7).

Section: выделение краевunclassified

Bio-inspired Neuron Model for Motion Detection on Base Signal Accumulation

Proceedings of the 32nd International Conference on Computer Graphics and Vision

Mangarakov

2022

“…Для выделения краев мы используем двухслойную нейронную сеть (рис. 2), предложенную нами в работе [7]. Выходы первого слоя такой сети помимо своей основной функции (выделения краев) служат также основой для работы нейронов конца линий.…”

Section: выделение краевunclassified

New End-Stopped Complex Cell Model Applicable in Convolutional Neural Networks

2021

jour

Self Cite

This article is dedicated to modeling the end-stopped neuron. This type of neuron gives the maximum response at the end of the line and is used to refine the edge. The article provides an overview of different models of end-stopped neurons. I have proposed a simpler and more accurate model of an end-stopped neuron based on the use of Gabor filters in antiphase. For this purpose, the models of simple and complex cells whose output is used in the proposed model are also described. Simple cells are based on the use of a Gabor filter, the parameters of which are also described in this article. The proposed model has shown its effectiveness.

Recognition of deaf gestures based on a bio-inspired neural network

Grif

2020

J. Phys.: Conf. Ser.

In this paper discusses the current situation in Russia and the world in the field of development of sign languages translation system. The main problems are formulated, and ways to solve them are given. One of the most important unresolved tasks is the task of recognizing the gestures of the deaf. To effectively solve it, an approach based on the development of bio-inspired neural networks is proposed. The architecture of a bio-inspired neural network, including four types of neurons, is described. New simpler MT neuron model proposed.