SF-Net: Structured Feature Network for Continuous Sign Language Recognition
It addresses continuous sign language recognition, a domain-specific problem for improving accessibility, but appears incremental as it builds on sequence-level supervision methods.
The paper tackles continuous sign language recognition by proposing SF-Net, which learns multiple semantic levels to address challenges like similar gestures and weak supervision, achieving clear outperformance over previous methods in accuracy and adaptability on two large-scale datasets.
Continuous sign language recognition (SLR) aims to translate a signing sequence into a sentence. It is very challenging as sign language is rich in vocabulary, while many among them contain similar gestures and motions. Moreover, it is weakly supervised as the alignment of signing glosses is not available. In this paper, we propose Structured Feature Network (SF-Net) to address these challenges by effectively learn multiple levels of semantic information in the data. The proposed SF-Net extracts features in a structured manner and gradually encodes information at the frame level, the gloss level and the sentence level into the feature representation. The proposed SF-Net can be trained end-to-end without the help of other models or pre-training. We tested the proposed SF-Net on two large scale public SLR datasets collected from different continuous SLR scenarios. Results show that the proposed SF-Net clearly outperforms previous sequence level supervision based methods in terms of both accuracy and adaptability.