Hard-wired visual filters for environment-agnostic object recognition

Abstract

Conventional deep neural networks (DNNs) are highly susceptible to variations in input domains, unlike biological brains which effectively adapt to environmental changes. Here, we demonstrate that hard-wired Gabor filters, replicating the structure of receptive fields in the brains early visual pathway, facilitate environment-agnostic object recognition without overfitting. Our approach involved fixing the pre-designed Gabor filters in the early layers of DNNs, preventing any alterations during training. Despite the restricted learning flexibility of this model, our networks maintained robust performance even under significant domain shifts, in contrast to conventional DNNs that typically fail in similar conditions. We found that our model effectively clustered identical "classes" across diverse domains, while conventional DNNs tend to cluster images by "domain" in the latent space. We observed that the fixed Gabor filters enabled networks to encode global shape information rather than local texture features, thereby mitigating the risk of overfitting. One sentence summaryHard-wired Gabor filters enable environment-agnostic object recognition without overfitting. Research HighlightsO_LIConventional deep neural networks (DNNs) are vulnerable to input domain variations C_LIO_LIHard-wired Gabor filters facilitate environment-agnostic object recognition C_LIO_LIFixed Gabor filters prevent overfitting and facilitate shape-based classifications C_LIO_LIOur model cluster identical "classes" while conventional DNNs cluster by "domain" C_LI