Performance Evaluation Of Yolov4-V8 On Indian Sign Language With Diverse Datasets
DOI:
https://doi.org/10.29284/qv93cg05Keywords:
Sign Language Recognition, Convolutional Neural Network, YOLO architecture, Data collections, Evaluation criteria.Abstract
Facilitating communication between individuals who are hearing and those who are deaf or mute presents a significant challenge. To bridge this gap, various systems have been developed to convert sign language into comprehensible forms by interpreting hand gestures into textual or spoken language. This research focuses on evaluating and contrasting multiple versions of the "You Only Look Once" (YOLO) algorithm in the context of Sign Language Recognition (SLR). YOLO, a high-speed and resource-efficient convolutional neural network (CNN) architecture, offers a promising approach for solving gesture recognition tasks. The study utilizes a range of datasets, including the ISL-CSLTR Dataset, FDMSE-ISL dataset, ISign Benchmark dataset, and the OpenHands dataset, all centered on Indian Sign Language (ISL). The comparative analysis is conducted using Python, assessing various evaluation parameters to determine the most suitable YOLO variant for sign recognition. This work aims to contribute to the development of a rapid and effective system for interpreting sign gestures.
References
[1] Cooper, H., et al. "Sign language recognition." Visual Analysis of Humans. Springer, London, 2011.
[2] Camgoz, N. C., Koller, O., Hadfield, S., & Bowden, R. (2018). "Neural sign language translation." CVPR, pp. 7784–7793.
[3] Ko, B., Kim, S., & Kim, C. (2019). "American sign language recognition using spatiotemporal features and 3D CNN." Sensors, 19(19), 4255.
[4] Saunders, B., et al. (2023). "OpenHands: Making Sign Language Recognition accessible with Computer Vision." arXiv preprint arXiv:2303.09538.
[5] Sahu, A., & Patel, V. (2021). "ISL Fingerspelling Recognition using CNN." International Journal of Computer Applications, 183(17), 1–6.
[6] Sharma, R., & Saini, M. (2022). "Real-time Indian Sign Language Recognition using Transfer Learning." Procedia Computer Science, 189, 234–240.
[7] Redmon, J., et al. (2016). "You Only Look Once: Unified, Real-Time Object Detection." CVPR, pp. 779–788.
[8] Bochkovskiy, A., Wang, C. Y., & Liao, H. Y. M. (2020). "YOLOv4: Optimal Speed and Accuracy of Object Detection." arXiv preprint arXiv:2004.10934.
[9] Raj, S., & Thomas, A. (2020). "Real-time ASL recognition using YOLOv3." International Journal of Advanced Computer Science and Applications, 11(4), 341–346.
[10] Patil, A., & Deshmukh, S. (2022). "Indian Sign Language Recognition using YOLOv4." Journal of Information Technology, 18(1), 45–52.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
