View source for XAI for yoga posture recognition

{{StudentProjectTemplate
|Summary=To analyze the classification of yoga postures and the potential misclassification due to occlusion and perspective of images
|TimeFrame=Spring 2025
|Supervisor=Cristofer Englund, Kevin Hernandez Diaz
|Level=Master
|Status=Open
}}
To analyze the classification of yoga postures and the potential misclassification due to occlusion and perspective of images, by exploring the feature space extracted from different classification models, to understand potential challenges or similarities between different postures. 

Networks that can be examined can be vision transformers, standard CNNs trained using different training strategies like siamese networks.