Exploitation of image statistics with sparse coding in the case of stereo vision

• Published in: Neural networks Vol. 135; pp. 158 - 176
• Main Authors: Ecke, Gerrit A., Papp, Harald M., Mallot, Hanspeter A.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.03.2021
• Subjects: Algorithms; Bayes Theorem; Compact code; Data Interpretation, Statistical; Efficient coding; Humans; Locally Competitive Algorithm (LCA); Pattern Recognition, Automated - methods; Probabilistic inference; Sparse coding; Stereo vision; Vision Disparity - physiology; Vision, Ocular - physiology; Visual Cortex - physiology; Visual Perception - physiology; Locally Competitive Algorithm (LCA); Stereo vision; Sparse coding; Efficient coding; Probabilistic inference; Compact code
• ISSN: 0893-6080; 1879-2782
• DOI: 10.1016/j.neunet.2020.12.016

The sparse coding algorithm has served as a model for early processing in mammalian vision. It has been assumed that the brain uses sparse coding to exploit statistical properties of the sensory stream. We hypothesize that sparse coding discovers patterns from the data set, which can be used to estimate a set of stimulus parameters by simple readout. In this study, we chose a model of stereo vision to test our hypothesis. We used the Locally Competitive Algorithm (LCA), followed by a naïve Bayes classifier, to infer stereo disparity. From the results we report three observations. First, disparity inference was successful with this naturalistic processing pipeline. Second, an expanded, highly redundant representation is required to robustly identify the input patterns. Third, the inference error can be predicted from the number of active coefficients in the LCA representation. We conclude that sparse coding can generate a suitable general representation for subsequent inference tasks. •Sparse coding discovers patterns from the dataset, which can be used to estimate a set of stimulus parameters by simple readout.•Successful stereo disparity estimation is possible with LCA sparse coding, followed by a naïve Bayes classifier.•An expanded, highly redundant representation is required to robustly identify the input patterns.•The inference error can be predicted from the number of active coefficients in the LCA representation.