Explicitly Modeling Pre-Cortical Vision with a Neuro-Inspired Front-End Improves CNN Robustness

TL;DR

This paper introduces biologically-inspired CNN front-end modules that simulate pre-cortical visual processing, significantly enhancing model robustness against various image corruptions while maintaining generalizability across architectures.

Contribution

The authors propose two novel CNN models with pre-cortical visual processing modules, demonstrating improved robustness over standard models across multiple corruption types.

Findings

RetinaNet achieves 12.3% robustness improvement.

EVNet achieves 18.5% robustness improvement.

Robustness gains generalize across different back-end architectures.

Abstract

While convolutional neural networks (CNNs) excel at clean image classification, they struggle to classify images corrupted with different common corruptions, limiting their real-world applicability. Recent work has shown that incorporating a CNN front-end block that simulates some features of the primate primary visual cortex (V1) can improve overall model robustness. Here, we expand on this approach by introducing two novel biologically-inspired CNN model families that incorporate a new front-end block designed to simulate pre-cortical visual processing. RetinaNet, a hybrid architecture containing the novel front-end followed by a standard CNN back-end, shows a relative robustness improvement of 12.3% when compared to the standard model; and EVNet, which further adds a V1 block after the pre-cortical front-end, shows a relative gain of 18.5%. The improvement in robustness was observed for all the different corruption categories, though accompanied by a small decrease in clean image accuracy, and generalized to a different back-end architecture. These findings show that simulating multiple stages of early visual processing in CNN early layers provides cumulative benefits for model robustness.