A multi-stage convolutional and self-attention architecture for high-precision sign language gesture recognition
Open Access
- 16 April 2026
- journal article
- research article
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 82 (6) , 355
Abstract
This paper presents a novel composite architecture for high-precision sign language gesture recognition referred to as ViTSLA. The model integrates convolutional inductive bias with transformer-based global attention to improve predictive accuracy and computational efficiency. Extensive experiments are conducted on three datasets using five-fold cross-validation, ablation analysis, and comparative evaluation against standard vision transformer (ViT) variants. Experimental results demonstrate consistently high classification accuracy exceeding 99.5% with minimal variance across folds, confirming robustness and balanced class-wise performance. Beyond predictive performance, this work addresses the computational demands of training and deploying hybrid deep architectures. Transformer-based attention mechanisms and high-resolution tokenization increase memory consumption and computational complexity. Therefore, efficient training requires parallel processing, GPU acceleration, and high-performance computing (HPC) environments to handle large-scale tensor operations and multi-fold cross-validation experiments. Furthermore, real-time inference capability is maintained at over 300 FPS, which is essential for deployment in latency-sensitive applications such as interactive systems and intelligent human–computer interfaces. Distributed computation further enables scalable training and stable convergence across datasets. Moreover, the low mean latency, stable P95 behavior, and high throughput confirm real-time deployment capability. These results demonstrate that the proposed architecture achieves high predictive performance while maintaining computational characteristics aligned with supercomputing-driven intelligent vision systems. The developed model achieves a favorable balance between accuracy, parameter efficiency, and computational cost. The results demonstrate that hybrid convolutional-attention architectures benefit from HPC-enabled parallelism, leading to scalable optimization and deployment-ready performance. These findings establish the relevance of the proposed approach within the domain of supercomputing-driven intelligent vision systems.Keywords
Funding Information
- Ankara Yıldırım Beyazıt Üniversitesi
This publication has 41 references indexed in Scilit:
- HCISLS: innovative sign language recognition system utilizing multi blocks convolutional methodUniversal Access in the Information Society, 2025
- Real-time arabic sign language recognition system using sensory glove and machine learningNeural Computing and Applications, 2025
- Real-time mobile application for Arabic sign alphabet recognition using pre-trained CNNSoft Computing, 2024
- Sign language interpretation using machine learning and artificial intelligenceNeural Computing and Applications, 2024
- Enhancing Indian sign language recognition through data augmentation and visual transformerNeural Computing and Applications, 2024
- Arabic sign language letters recognition using Vision TransformerMultimedia Tools and Applications, 2024
- Fostering inclusivity through effective communication: Real-time sign language to speech conversion system for the deaf and hard-of-hearing communityMultimedia Tools and Applications, 2023
- Best practices for sign language technology researchUniversal Access in the Information Society, 2023
- Implementation of wearable glove for sign language expression based on deep learningMicrosystem Technologies, 2023
- Regularizing self-attention on vision transformers with 2D spatial distance lossArtificial Life and Robotics, 2022