Recurrent Feedback Improves Recognition of Partially Occluded Objects

TL;DR

This paper demonstrates that recurrent neural network architectures significantly improve object recognition accuracy under occlusion, outperforming feedforward models and effectively revising initial guesses for challenging stimuli.

Contribution

It provides evidence that recurrence enhances recognition of occluded objects in artificial neural networks, supported by novel datasets and comparative analysis.

Findings

Recurrent models outperform feedforward models in occlusion recognition.

Recurrent feedback can revise initial guesses for challenging stimuli.

Recurrent connectivity improves accuracy under occlusion conditions.

Abstract

Recurrent connectivity in the visual cortex is believed to aid object recognition for challenging conditions such as occlusion. Here we investigate if and how artificial neural networks also benefit from recurrence. We compare architectures composed of bottom-up, lateral and top-down connections and evaluate their performance using two novel stereoscopic occluded object datasets. We find that classification accuracy is significantly higher for recurrent models when compared to feedforward models of matched parametric complexity. Additionally we show that for challenging stimuli, the recurrent feedback is able to correctly revise the initial feedforward guess.