ColorFoil: Investigating Color Blindness in Large Vision and Language Models

TL;DR

This paper introduces ColorFoil, a benchmark to evaluate large vision and language models' ability to perceive colors, revealing significant gaps in their robustness and color discrimination capabilities in zero-shot settings.

Contribution

The paper presents a new benchmark, ColorFoil, for assessing color perception in V&L models and evaluates seven models, highlighting their strengths and weaknesses in color recognition.

Findings

ViLT and BridgeTower outperform others in color perception.

CLIP-based models and GroupViT struggle with distinct color differentiation.

Models show limited robustness in complex linguistic and visual attribute understanding.

Abstract

With the utilization of Transformer architecture, large Vision and Language (V&L) models have shown promising performance in even zero-shot settings. Several studies, however, indicate a lack of robustness of the models when dealing with complex linguistics and visual attributes. In this work, we introduce a novel V&L benchmark - ColorFoil, by creating color-related foils to assess the models' perception ability to detect colors like red, white, green, etc. We evaluate seven state-of-the-art V&L models including CLIP, ViLT, GroupViT, and BridgeTower, etc. in a zero-shot setting and present intriguing findings from the V&L models. The experimental evaluation indicates that ViLT and BridgeTower demonstrate much better color perception capabilities compared to CLIP and its variants and GroupViT. Moreover, CLIP-based models and GroupViT struggle to distinguish colors that are visually distinct to humans with normal color perception ability.