MTVQA: Benchmarking Multilingual Text-Centric Visual Question Answering. A comprehensive evaluation of multimodal large model multilingual text perception and comprehension capabilities across nine widely-used yet low-resource languages. - View it on GitHub