We present a novel approach for automating character recognition and dialogue extraction in comics for dubbing applications. Our system operates in four main phases: character detection with gender identification, speech balloon localization with text extraction, text correction, and voice dubbing. We leverage YOLOv11 for object detection and OCR for text retrieval, followed by transformer-based text correction. This paper focuses on the first two phases, presenting effective solutions for extracting characters, dialog balloon and their associated dialogues. Our experimental results demonstrate the practical applicability of the proposed methodology with high accuracy rates in character-dialogue mapping for the Doraemon comic series. We have obtained positive results that can be applied in practice.