Leveraging Human Attention in Novel Object Captioning [conference paper]
Conference
30th International Joint Conference on Artificial Intelligence (IJCAI) - August 19 - 26, 2021
Authors
Xianyu Chen (Ph.D. student), Ming Jiang (postdoctoral associate), Qi Zhao (associate professor)
Abstract
Image captioning models depend on training with paired image-text corpora, which poses various challenges in describing images containing novel objects absent from the training data. While previous novel object captioning methods rely on external image taggers or object detectors to describe novel objects, we present the Attention based Novel Object Captioner (ANOC) that complements novel object captioners with human attention features that characterize generally important information independent of tasks. It introduces a gating mechanism that adaptively incorporates human attention with self-learned machine attention, with a Constrained Self-Critical Sequence Training method to address the exposure bias while maintaining constraints of novel object descriptions. Extensive experiments conducted on the no caps and Held-Out COCO datasets demonstrate that our method considerably outperforms the state of-the-art novel object captioners.
Link to full paper
Leveraging Human Attention in Novel Object Captioning
Keywords
computer vision, artificial intelligence