Investigating the Gestalt Principle of Closure in Deep Convolutional Neural Networks

TL;DR

This paper explores whether deep convolutional neural networks perceive objects similarly to humans by testing their ability to recognize incomplete shapes based on the Gestalt principle of closure, revealing reliance on complete edges.

Contribution

It introduces a protocol to assess closure perception in CNNs and evaluates well-known models on incomplete shape classification, highlighting their limitations.

Findings

Performance decreases with more edge removal

Current CNNs rely heavily on complete edge information

Models struggle with incomplete shape recognition

Abstract

Deep neural networks perform well in object recognition, but do they perceive objects like humans? This study investigates the Gestalt principle of closure in convolutional neural networks. We propose a protocol to identify closure and conduct experiments using simple visual stimuli with progressively removed edge sections. We evaluate well-known networks on their ability to classify incomplete polygons. Our findings reveal a performance degradation as the edge removal percentage increases, indicating that current models heavily rely on complete edge information for accurate classification. The data used in our study is available on Github.