Enhancing Continuous Sign Language Recognition through MGPT-based Segmentation and Structured Position-Aware Decoding

RAIDAL: Redundancy-Aware Information Density Active Learning for CTC-Based Continuous Sign Language Recognition

This work repurposes the CTC decoder to restrict representation-based scoring to decoder-aligned gloss regions, rather than exposing the acquisition function to the entire unfiltered video, and introduces RAIDAL, which achieves its strongest data-efficiency gains over competing baselines in large-vocabulary, budget-limited settings, while remaining competitive in the smaller-vocabulary, large-budget setting.

TPA-Seq2Seq: Tri-Prior Aligned Sequence-to-Sequence Learning for Continuous Sign Language Recognition

With the growing emphasis on accessibility-oriented technologies and inclusive intelligent systems, Continuous Sign Language Recognition (CSLR) has attracted increasing attention as a key technique for bridging communication between Deaf and hearing communities. However, existing methods still suffer from insufficient exploitation of visual information, weak temporal alignment supervision, and inconsistency between training and inference, making it difficult to jointly improve recognition accuracy and model robustness. To address these issues, we propose TPA-Seq2Seq (Tri-Prior Aligned Seq2Seq), a tri-prior enhanced Seq2Seq framework for character-level CSLR. Specifically, we introduce PGF (Pose-Guided Fusion) to extract hand-arm-face keypoint descriptors offline through the MediaPipe pipeline and fuse them with RGB-based temporal semantics in a lightweight manner, thereby supplementing fine-grained visual priors. We further design ATAL (Auxiliary Temporal Alignment Loss), a CTC-based auxiliary constraint imposed on the encoder side to strengthen explicit temporal alignment supervision. In addition, we propose SSF (Scheduled Semantic Forcing), a piecewise teacher-forcing decay strategy that alleviates exposure bias and improves generalization. Experiments on the CSL dataset show that TPA-Seq2Seq reduces WER compared with the ResNet18-LSTM baseline and maintains consistent performance across different random seeds.

SignDino: Self-Supervised Sign Language Representation Learning via Temporal-Axis Self-Distillation

Self-supervised sign language representation learning must model two properties not central to natural-image SSL: signs are produced by a small set of anatomically distinct articulators, and their meaning depends on the temporal organisation of those articulators. We introduce SignDino, a self-supervised sign-video encoder that moves the DINOv3 student--teacher recipe from the spatial domain of image crops to the temporal domain of tracked sign streams. Each video is decomposed into left-hand, right-hand, and face streams by a detector-first YOLOv8n+ByteTrack pipeline. A frozen DINOv3 ViT-B/16 embeds each per-frame anatomical crop, while lightweight temporal Transformers, not the image backbone, form the student and EMA teacher. They are trained by temporal DINO self-distillation, frame-level masked-token prediction in the style of iBOT, KoLeo feature spreading, and Gram anchoring of the frame-to-frame similarity structure. This design keeps strong image-level visual primitives fixed and learns only how articulator states evolve across time. We evaluate on sign-to-English translation, isolated sign recognition, and fingerspelling detection benchmarks. Across these tasks, SignDino provides a strong public self-supervised representation and shows competitive or state-of-the-art performance under matched downstream evaluation.

A landmark-based temporal pipeline for real-time adjective recognition in Indian sign language

Dynamic Sign Language Recognition: A Hybrid Approach Combining MediaPipe and LSTM

Supercell thunderstorm paper publishing-based optimization enabled deep learning for sign language recognition using videos

Sign language is an essential communication system for hearing-impaired individuals, which mainly depends upon complex hand gestures and facial expressions. Automating Sign Language Recognition (SLR) from videos can enhance barrier-free communication, yet it remains challenging due to the subtle nature of signs in diverse environments. Existing modules often need extensive manual feature extraction and struggle with real-time applications due to high latency. Thus, an effective model for SLR using videos is presented, named Supercell Thunderstorm Paper Publishing-based Optimization enabled Convolutional Grid Long Short-Term Memory (STPPO_CGLSTM). The frames are extracted from the input video. Then, the Arithmetic Mean Filter is employed for pre-processing the extracted frames. Moreover, humans are segmented by employing correlational spectral clustering. Thereafter, hand action unit detection is carried out by considering the AU-Net. Following this detection task, effectual features are extracted. Lastly, sign language is recognized by utilizing CGLSTM, where the hyperparameters of CGLSTM are trained using STPPO. The analytical measures, namely, accuracy, Positive Predictive Value (PPV), Negative Predictive Value (NPV), and False Omission Rate (FOR), obtained better outcomes for STPPO_CGLSTM, which is 96.707%, 98.219%, 94.016% and 5.984% using k-fold cross validation.