Hidden in Plain Sight: Evaluating Abstract Shape Recognition in Vision-Language Models

TL;DR

This paper evaluates the ability of current vision-language models to recognize abstract shapes, revealing significant limitations despite their apparent reliance on shape features, and introduces IllusionBench to quantify these challenges.

Contribution

The paper introduces IllusionBench, a novel dataset designed to test shape recognition in vision-language models, highlighting their ongoing struggles with abstract shape perception.

Findings

VLMs perform poorly on shape recognition tasks compared to humans.

Current VLMs still rely on spurious features rather than true shape understanding.

IllusionBench effectively exposes the limitations of VLMs in perceiving abstract shapes.

Abstract

Despite the importance of shape perception in human vision, early neural image classifiers relied less on shape information for object recognition than other (often spurious) features. While recent research suggests that current large Vision-Language Models (VLMs) exhibit more reliance on shape, we find them to still be seriously limited in this regard. To quantify such limitations, we introduce IllusionBench, a dataset that challenges current cutting-edge VLMs to decipher shape information when the shape is represented by an arrangement of visual elements in a scene. Our extensive evaluations reveal that, while these shapes are easily detectable by human annotators, current VLMs struggle to recognize them, indicating important avenues for future work in developing more robust visual perception systems. The full dataset and codebase are available at: \url{https://arshiahemmat.github.io/illusionbench/}