MUNIChus: Multilingual News Image Captioning Benchmark
This addresses a data scarcity problem for researchers working on multilingual news image captioning, though it is incremental as it primarily provides a new benchmark.
The authors tackled the lack of multilingual datasets in news image captioning by creating MUNIChus, the first benchmark covering 9 languages including low-resource ones, and found that state-of-the-art models still struggle with this task.
The goal of news image captioning is to generate captions by integrating news article content with corresponding images, highlighting the relationship between textual context and visual elements. The majority of research on news image captioning focuses on English, primarily because datasets in other languages are scarce. To address this limitation, we create the first multilingual news image captioning benchmark, MUNIChus, comprising 9 languages, including several low-resource languages such as Sinhala and Urdu. We evaluate various state-of-the-art neural news image captioning models on MUNIChus and find that news image captioning remains challenging. We also make MUNIChus publicly available with over 20 models already benchmarked. MUNIChus opens new avenues for further advancements in developing and evaluating multilingual news image captioning models.