Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{185351,
        author = {Vinay Pandey and Priya koshle},
        title = {Optimizing Sign Language Recognition through Hybrid Machine Learning Architectures},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {5},
        pages = {1124-1133},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=185351},
        abstract = {We propose a hybrid machine learning approach to the efficient SLR by combining CNN with Long Short-Term Memory (LSTM) and Transformer architectures. The model can effectively learn spatial and temporal information from gesture sequences. Preprocessing: Grayscale conversion, Gaussian filtering, hand segmentation and data augmentation preprocess raw input data to improve quality and robustness of the model. Experimental results of the hybrids on both custom and standard benchmarks showed a significant superiority over conventional approaches. CNN–Transformer architecture obtains the best performance of 94.1%, and is closely followed by CNN–LSTM with an accuracy of 92.6%, indicating that our method has better performance on dynamic gesture recognition. Results show that the proposed hybrid architecture has better performance in terms of recognition accuracy, temporal consistency and computational efficiency, which is more suitable for real-time and friendly communication systems for deaf-mute.},
        keywords = {Sign Language Recognition, Hybrid Machine Learning, Deep Learning, CNN, RNN, LSTM, Gesture Recognition, Human-Computer Interaction, Accessibility, Real-Time Recognition.},
        month = {October},
        }