Multi-Stream Keypoint Attention Network for Sign Language Recognition and Translation

TL;DR

The paper tackles robust sign language recognition and translation by shifting to a four-stream keypoint-based approach that models interactions among hands, face, and body. It introduces MSKA-SLR with decoupled keypoint streams and multi-stream attention, augmented by spatial-global regularization and frame-level self-distillation, and extends it to MSKA-SLT using a translation network (mBART). The method achieves strong results on Phoenix-2014, Phoenix-2014T, and CSL-Daily, including state-of-the-art translation on Phoenix-2014T, while maintaining background-robustness and computational efficiency. These contributions offer a practical, interpretable baseline for keypoint-driven SLR/SLT research with a focus on cross-stream interactions and knowledge transfer between streams.

Abstract

Sign language serves as a non-vocal means of communication, transmitting information and significance through gestures, facial expressions, and bodily movements. The majority of current approaches for sign language recognition (SLR) and translation rely on RGB video inputs, which are vulnerable to fluctuations in the background. Employing a keypoint-based strategy not only mitigates the effects of background alterations but also substantially diminishes the computational demands of the model. Nevertheless, contemporary keypoint-based methodologies fail to fully harness the implicit knowledge embedded in keypoint sequences. To tackle this challenge, our inspiration is derived from the human cognition mechanism, which discerns sign language by analyzing the interplay between gesture configurations and supplementary elements. We propose a multi-stream keypoint attention network to depict a sequence of keypoints produced by a readily available keypoint estimator. In order to facilitate interaction across multiple streams, we investigate diverse methodologies such as keypoint fusion strategies, head fusion, and self-distillation. The resulting framework is denoted as MSKA-SLR, which is expanded into a sign language translation (SLT) model through the straightforward addition of an extra translation network. We carry out comprehensive experiments on well-known benchmarks like Phoenix-2014, Phoenix-2014T, and CSL-Daily to showcase the efficacy of our methodology. Notably, we have attained a novel state-of-the-art performance in the sign language translation task of Phoenix-2014T. The code and models can be accessed at: https://github.com/sutwangyan/MSKA.

Figures (3)

• Figure 1: (a)We choose a sign language video from the Phoenix-2014T dataset and display its gloss sequence alongside the corresponding text. The objective of sign language recognition (SLR) is to instruct models in producing matching gloss representations derived from sign language videos. Conversely, the task of sign language translation (SLT) entails creating textual representations that align with sign language videos. (b) MSKA-SLT is constructed on the foundation of MSKA-SLR to facilitate SLT. Keypoint sequences are depicted in coordinate form.
• Figure 2: The overview of our MSKA-SLR. The whole network is jointly supervised by the CTC losses and the self-distillation losses. Keypoints are represented in coordinate format.
• Figure 3: Illustration of the attention module. We show the body attention module as an example. The others attention module is an analogy. The green rounded rectangle box represents a single-head self-attention module. There are totally $h$ self-attention heads, whose output are concatenated and fed into two linear layers to obtain the output. LReLU represents the leaky ReLU maas2013rectifier.