Inspiration
My inspiration comes from my research done at the University of Washington on computer vision. I wanted to apply a fine-tuned DL model to provide a more affordable and ubiquitous way of using computer vision and AR environments.
Infrastructure
The code runs in 3 main parts. A fined tuned DL model from Open CV with further training on scraped video data of athletics events is used to track and map a user or video. The data points from this tracking is then sent over a socket webserver to a unity game environment where a digital simulacrum of you is represented and mirrored.
Use Cases
Like any AR/VR project, the use cases of this is endless. From gaming to medicine, this tech can be used extensively for those who are thrill-seeking athletes cooped in a room to depressed med students who can preform digital surgery as practice. The only thing limiting the ability for further expansion is the general weakness of not being immersive enough and the limitations of camera technologies.
Log in or sign up for Devpost to join the conversation.