DIMT25@ICDAR2025: HW-TSC's End-to-End Document Image Machine Translation System Leveraging Large Vision-Language Model
This work provides an incremental improvement for researchers and practitioners in document analysis and machine translation by enhancing translation capabilities for complex layouts.
The paper tackled the problem of end-to-end document image machine translation for complex layouts by introducing a training framework that combines multi-task learning with perceptual chain-of-thought using a large vision-language model, resulting in a system that addresses both OCR-based and OCR-free tasks within a unified framework.
This paper presents the technical solution proposed by Huawei Translation Service Center (HW-TSC) for the "End-to-End Document Image Machine Translation for Complex Layouts" competition at the 19th International Conference on Document Analysis and Recognition (DIMT25@ICDAR2025). Leveraging state-of-the-art open-source large vision-language model (LVLM), we introduce a training framework that combines multi-task learning with perceptual chain-of-thought to develop a comprehensive end-to-end document translation system. During the inference phase, we apply minimum Bayesian decoding and post-processing strategies to further enhance the system's translation capabilities. Our solution uniquely addresses both OCR-based and OCR-free document image translation tasks within a unified framework. This paper systematically details the training methods, inference strategies, LVLM base models, training data, experimental setups, and results, demonstrating an effective approach to document image machine translation.