SimulSLT: End-to-End Simultaneous Sign Language Translation
This addresses the problem of real-time application limitations for sign language translation users, representing a novel method for a known bottleneck.
The paper tackles the high inference latency in sign language translation by proposing SimulSLT, the first end-to-end simultaneous model that translates videos concurrently, achieving higher BLEU scores than non-simultaneous models on the RWTH-PHOENIX-Weather 2014T dataset while maintaining low latency.
Sign language translation as a kind of technology with profound social significance has attracted growing researchers' interest in recent years. However, the existing sign language translation methods need to read all the videos before starting the translation, which leads to a high inference latency and also limits their application in real-life scenarios. To solve this problem, we propose SimulSLT, the first end-to-end simultaneous sign language translation model, which can translate sign language videos into target text concurrently. SimulSLT is composed of a text decoder, a boundary predictor, and a masked encoder. We 1) use the wait-k strategy for simultaneous translation. 2) design a novel boundary predictor based on the integrate-and-fire module to output the gloss boundary, which is used to model the correspondence between the sign language video and the gloss. 3) propose an innovative re-encode method to help the model obtain more abundant contextual information, which allows the existing video features to interact fully. The experimental results conducted on the RWTH-PHOENIX-Weather 2014T dataset show that SimulSLT achieves BLEU scores that exceed the latest end-to-end non-simultaneous sign language translation model while maintaining low latency, which proves the effectiveness of our method.