{"slug": "transslr-a-lightweight-transformer-for-sign-language-recognition", "title": "TransSLR: A Lightweight Transformer for Sign Language Recognition", "summary": "Researchers propose TransSLR, a lightweight Temporal Transformer Encoder trained from scratch on 64-frame normalized pose sequences, achieving 80.39% accuracy on the CASL-W60 benchmark for Central African Sign Language recognition, surpassing the prior best of 69.93% by 10.46 percentage points. The model operates on geometric keypoint representations rather than raw RGB, enabling signer-independent generalization and reduced computational overhead for resource-constrained deployment.", "body_md": "arXiv:2608.06407v1 Announce Type: new\nAbstract: Automated Sign Language Recognition for under-represented languages remains a largely unsolved problem. Central African Sign Language (CASL) exemplifies this gap: the only available bench-mark, CASL-W60, has a best reported accuracy of 69.93%, and we show that the common heuristic of fine-tuning high-resource models fails to close it. This failure stems from two compounding factors: the limited scale of available CASL data and the significant lexical and visual domain gap between CASL and large-scale corpora such as WLASL, which renders pre-trained representations largely uninformative.\nTo address this, we propose TransSLR, a lightweight Temporal Transformer Encoder trained from scratch on 64-frame normalized pose sequences, with average pooling and a classification head. By operating on geometric keypoint representations rather than raw RGB, TransSLR achieves signer-independent generalization without relying on visual appearance. On the CASL-W60 benchmark, TransSLR establishes a new state-of-the-art accuracy of 80.39%, surpassing the prior best by +10.46%. Beyond accuracy, our encoder-only design significantly reduces computational overhead, making deployment feasible in resource-constrained environments. We conduct extensive experiments on the CASL-W60 benchmark, comparing against RGB-based and multimodal baselines, and demonstrate that TransSLR achieves state-of-the-art performance.", "url": "https://wpnews.pro/news/transslr-a-lightweight-transformer-for-sign-language-recognition", "canonical_source": "https://arxiv.org/abs/2608.06407", "published_at": "2026-08-10 04:00:00+00:00", "updated_at": "2026-08-10 04:14:16.805761+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "computer-vision"], "entities": ["TransSLR", "Central African Sign Language (CASL)", "CASL-W60", "WLASL"], "alternates": {"html": "https://wpnews.pro/news/transslr-a-lightweight-transformer-for-sign-language-recognition", "markdown": "https://wpnews.pro/news/transslr-a-lightweight-transformer-for-sign-language-recognition.md", "text": "https://wpnews.pro/news/transslr-a-lightweight-transformer-for-sign-language-recognition.txt", "jsonld": "https://wpnews.pro/news/transslr-a-lightweight-transformer-for-sign-language-recognition.jsonld"}}