Fusion of AI Mocap Technology in Augmenting the Human Motion
Youtube Link : https://www.youtube.com/watch?v=WEnoolpvbcU&t=82s
The recognition of actions holds a pivotal role across a spectrum of applications. This research works puts forth a novel approach aimed at discerning human actions by harnessing the wealth of data derived from motion capture actions.
The goal of this research work is to test the hypothesis of general and global motion fingerprints for both activity classification and individual identification. We propose a comprehensive and self-sustainable system dubbed as FamtamAI: Fusion of Ai-Mocap Technology in Augmenting the human Motion. The proposed model is able to produce fingerprints regards to individual atomic motion and can leverage this knowledge to enhance security, ergonomic conditions, recommend improvements, and optimize performance in sports and other activities.
Our main contribution to the domain is,
- Proposed a novel idea of Global Motion Fingerprint for both activity classification and identity recognition.
- Proposed a model "Famtam AI" for recognizing the global motion fingerprint and tested it under the standard protocol.
- Contributed to open-source motion projects including bvh-python Github repository.
We tested our model with Berkerly MHAD test dataset and achieved an accuracy of circa 98% and compared the results against various other existing researchers.
Keywords: AI, Famtam AI, Motion Fingerprint, Motion Capture, Motion Analysis, Identity Recognition, Activity Recognition