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{167821,
        author = {Sumon Singh and Deepali Handgar},
        title = {Yoga pose classification from images using transfer learning approach},
        journal = {International Journal of Innovative Research in Technology},
        year = {2024},
        volume = {11},
        number = {4},
        pages = {650-657},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=167821},
        abstract = {Yoga is a practice that originated in ancient India. Hatha Yoga is a type of physical activity that consists of postures (also called ‘asanas’) in continuous sequence along with regulated breathing. Yoga is helpful to balance our mind and body through meditation, exercise, and breathing work. Various studies show yoga’s benefits for arthritis, mental health, women’s health, and other specialties. For all these reasons, yoga has seen immense popularity throughout the world. It’s important to correctly identify the asana a person needs to perform according to his/her needs. This paper uses a transfer learning approach to identify the yoga pose shown in each picture. A total of 1551 images of 5 different yoga postures were used and resized for ease of computing. 10 models were used for the classification of the images and their evaluation metrics were compared to see which model gave better results. The models used were VGG16, VGG19, InceptionV3, DenseNet201, ResNet50V2, ResNet152V2, ResNet101V2, MobileNet, MobileNetV2 and InceptionResNetV2. Out of all these models, VGG16 outperforms and the validation accuracy witnessed is 94.47% (with un-augmented data).},
        keywords = {Deep learning, Image classification, Transfer learning, Yoga pose classification},
        month = {September},
        }