Vadim Glezer’s modules model as a possible basis for vision perception description
- 1 August 2024
- journal article
- Published by Optica Publishing Group in Journal of Optical Technology
- Vol. 91 (8) , 570-576
Abstract
Subject of study. Mechanisms of visual perception were studied. Aim of study. The aim of this study was to determine the feasibility of using the modules model to describe visual perception. This model was proposed by V. D. Glezer based on electrophysiological data. Method. Psychophysical methods and modeling were employed. Experiments on image recognition and the estimation of spatial interval and circle sizes in the Delboeuf and Ebbinghaus illusions were conducted. The experimental data were fitted using two models: one based on full image spectra and another based on the modules model, in which images were decomposed into finite Fourier series. For each image, a module of the optimal size was selected to retain the maximum energy of the image. Two combinations of such modules were analyzed. Main results. The results showed that the optimal module size depended on the image. Errors in image recognition and similarity assessments correlated with the distances between images calculated in the feature space as the norm of the difference in their spectra. These errors also partially correlated with the distances obtained using one variant of the modules model. The modules model adequately approximated the segmentation and size estimation data, as confirmed by the analysis of paintings. Thus, the modules model can describe the mechanisms of size estimation and segmentation. However, for image recognition, the model requires improvement, specifically in terms of the introduction of interactions between modules of different sizes. Thus, the modules model serves as a preliminary approximation for describing visual perception. Practical significance. The modules model may be applied to image analysis and the development of artificial neural networks for segmentation and object recognition.Keywords
Funding Information
- Russian Academy of Sciences (1021062411653-4-3.1.8)
This publication has 25 references indexed in Scilit:
- Convolutional Neural Networks: An Overview and Its Applications in Pattern Recognition2020
- An Analysis Of Convolutional Neural Networks For Image ClassificationProcedia Computer Science, 2018
- Size Estimates in Ebbinghaus Illusion in Adults and Children of Different AgeHuman Physiology, 2004
- Peripheral and central delay in processing high spatial frequencies: reaction time and VEP latency studiesVision Research, 1999
- Harmonic basis functions for spatial coding in the cat striate cortexVisual Neuroscience, 1989
- Identification confusions among letters of the alphabet.Journal of Experimental Psychology: Human Perception and Performance, 1984
- Fast filter transform for image processingComputer Graphics and Image Processing, 1981
- Linear and non-linear properties of simple and complex receptive fields in area 17 of the cat visual cortexBiological Cybernetics, 1980
- On the existence of neurones in the human visual system selectively sensitive to the orientation and size of retinal imagesThe Journal of Physiology, 1969
- Application of fourier analysis to the visibility of gratingsThe Journal of Physiology, 1968