Copyright © 2025 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{182990,
        author = {Remya PK and Rajkumar KK},
        title = {Enhanced Hand Segmentation Using UNet: A Comparative Study with Classical Approaches},
        journal = {International Journal of Innovative Research in Technology},
        year = {},
        volume = {12},
        number = {no},
        pages = {99-105},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=182990},
        abstract = {The segmentation of hand gestures from complex backgrounds is essential for accurately classifying classical dance movements, as it enhances feature extraction and recognition performance. This paper presents a robust segmentation framework based on the UNet architecture, a deep learning model well-suited for pixel-wise segmentation. Initially, various traditional segmentation techniques are analyzed and evaluated; among them, UNet demonstrates superior performance in handling the complexity of classical dance gestures. The proposed method consists of two main stages: developing a UNet-based model for hand segmentation and refining it to capture the intricate and expressive movements characteristic of classical dance. The model is trained on a carefully curated dataset featuring diverse hand shapes, skin tones, and poses to ensure strong generalization. Post-segmentation, feature extraction is optimized to focus on the most relevant elements for classification. Robustness and adaptability are further improved using data augmentation and transfer learning techniques. Experimental results show that the proposed system surpasses existing methods in accuracy, real-time performance, and resilience under varying conditions. Applications include real-time gesture recognition in classical dance training, augmented reality, and enhanced human-computer interaction.},
        keywords = {Hand Segmentation, Deep Learning, UNet, Gesture Recognition, Mudra Classification},
        month = {},
        }