Decoding Emotions in Abstract Art: Cognitive Plausibility of CLIP in Recognizing Color-Emotion Associations

TL;DR

This paper evaluates CLIP's ability to recognize emotions in abstract art, revealing its partial success and highlighting differences from human emotional processing, especially in color-emotion associations.

Contribution

It provides a comprehensive analysis of CLIP's performance in emotion recognition in abstract art and explores color-emotion interactions, revealing cognitive plausibility limitations.

Findings

CLIP achieves above baseline but low accuracy in emotion recognition.

CLIP shows stronger color-emotion associations than humans.

Disparity exists between human and machine emotion processing.

Abstract

This study investigates the cognitive plausibility of a pretrained multimodal model, CLIP, in recognizing emotions evoked by abstract visual art. We employ a dataset comprising images with associated emotion labels and textual rationales of these labels provided by human annotators. We perform linguistic analyses of rationales, zero-shot emotion classification of images and rationales, apply similarity-based prediction of emotion, and investigate color-emotion associations. The relatively low, yet above baseline, accuracy in recognizing emotion for abstract images and rationales suggests that CLIP decodes emotional complexities in a manner not well aligned with human cognitive processes. Furthermore, we explore color-emotion interactions in images and rationales. Expected color-emotion associations, such as red relating to anger, are identified in images and texts annotated with emotion labels by both humans and CLIP, with the latter showing even stronger interactions. Our results highlight the disparity between human processing and machine processing when connecting image features and emotions.