ComiCap: A VLMs pipeline for dense captioning of Comic Panels

TL;DR

This paper introduces ComiCap, a pipeline utilizing Vision-Language Models to generate dense, grounded captions for comic panels, achieving superior results without additional training and enabling large-scale annotation of comic books.

Contribution

It presents a novel VLM-based pipeline for dense captioning of comic panels, including an attribute-retaining metric and a new annotated dataset for evaluation.

Findings

The pipeline outperforms specialized models in caption quality.

It can annotate over 2 million comic panels efficiently.

The method requires no additional training of VLMs.

Abstract

The comic domain is rapidly advancing with the development of single- and multi-page analysis and synthesis models. Recent benchmarks and datasets have been introduced to support and assess models' capabilities in tasks such as detection (panels, characters, text), linking (character re-identification and speaker identification), and analysis of comic elements (e.g., dialog transcription). However, to provide a comprehensive understanding of the storyline, a model must not only extract elements but also understand their relationships and generate highly informative captions. In this work, we propose a pipeline that leverages Vision-Language Models (VLMs) to obtain dense, grounded captions. To construct our pipeline, we introduce an attribute-retaining metric that assesses whether all important attributes are identified in the caption. Additionally, we created a densely annotated test set to fairly evaluate open-source VLMs and select the best captioning model according to our metric. Our pipeline generates dense captions with bounding boxes that are quantitatively and qualitatively superior to those produced by specifically trained models, without requiring any additional training. Using this pipeline, we annotated over 2 million panels across 13,000 books, which will be available on the project page https://github.com/emanuelevivoli/ComiCap.