1 paper · 1 filter
Shuitsu Koyama, Kazuki Matsuda, Yuiga Wada +3
Automatic evaluation of image and video captioning is essential for benchmarking multimodal systems, although standard evaluation metrics show limited alignment with human judgment…