Re:Cognize -- Open-Set Comic Character Re-Identification

2026-09-27

AI Summary

This study addresses the open-set character re-identification challenge in streaming manga reading, revealing that the performance bottleneck when models autonomously construct character galleries stems from erroneous match acceptance rather than insufficient visual perception. To overcome this, we propose ReCast, a framework incorporating a page-context-aware self-verification mechanism and a parameter-free learning strategy for contrastive criteria. This approach dynamically updates the gallery only when new matches significantly outperform existing entries, combined with moving-average feature aggregation to ensure high-quality gallery growth. Without requiring predefined rosters or manual annotations, the proposed method recovers one-third to two-thirds of the Top-1 accuracy achievable under ideal labeling conditions, substantially improving open-set sequential retrieval performance.

Abstract

A manga reader meets a character on one page and knows them on sight a hundred pages later, without ever being handed a cast list. Re-identifying comic characters demands the same, open-set and sequential: pages arrive as a stream in reading order, new faces appear before anyone names them, and the cast is assembled as the story is read. $\textbf{Re:Cognize}$ evaluates recognition as the story is read, not against a cast handed over in advance: four protocols on one query stream, from closed-set retrieval to a cast the model must build and grow itself. The surprise is where models fail. Recognising is close to solved: one reference image per character already ranks as well as a gallery built in advance. Knowing what to believe is not: a model that adds its own matches makes its cast worse, while the same growth with correct labels would gain over twenty points of top-1 accuracy. The bottleneck is acceptance, not vision, and one comparison decides it: an addition pays exactly when it is right more often than the cast already was on the queries it takes over. The comparison has nothing to fit, and measured on half of a new corpus it calls the other half correctly. $\textbf{ReCast}$ puts it to work with nothing fitted on data: a cast sheet of one running average per character, grown only where the page itself vouches for a crop. It recovers a third to two thirds of what perfect labels would, depending on whether the cast starts from random examples or from first appearances. Re:Cognize measures whether a model can read along; ReCast is a cast that does. Our claims are on identity maintenance, recognising characters already met; the emergence of new ones is measured as a diagnostic under a fixed reference rule, and we propose no method for it.

Problem

• Open-Set Re-Identification
• Comic Character Recognition
• Manga
• Identity Maintenance
• Sequential Recognition

Innovation

• Open-Set Re-Identification
• Comic Character Recognition
• Dynamic Cast Building
• ReCast
• Acceptance Strategy