Zero-Shot Character Identification and Speaker Prediction in Comics via Iterative Multimodal Fusion

TL;DR

This paper introduces a zero-shot multimodal framework for identifying characters and predicting speakers in comics without requiring annotations, leveraging large language models and iterative fusion techniques.

Contribution

It presents the first multimodal, zero-shot approach for character and speaker recognition in comics, eliminating the need for comic-specific training data.

Findings

Effective zero-shot character identification and speaker prediction

Establishes a robust baseline for multimodal comic analysis

Applicable to any comic series without additional training

Abstract

Recognizing characters and predicting speakers of dialogue are critical for comic processing tasks, such as voice generation or translation. However, because characters vary by comic title, supervised learning approaches like training character classifiers which require specific annotations for each comic title are infeasible. This motivates us to propose a novel zero-shot approach, allowing machines to identify characters and predict speaker names based solely on unannotated comic images. In spite of their importance in real-world applications, these task have largely remained unexplored due to challenges in story comprehension and multimodal integration. Recent large language models (LLMs) have shown great capability for text understanding and reasoning, while their application to multimodal content analysis is still an open problem. To address this problem, we propose an iterative multimodal framework, the first to employ multimodal information for both character identification and speaker prediction tasks. Our experiments demonstrate the effectiveness of the proposed framework, establishing a robust baseline for these tasks. Furthermore, since our method requires no training data or annotations, it can be used as-is on any comic series.