I please need a golf club analysis application developed based on computer vision. Club movement in a video needs to be precisely tracked, graphically reconstructed for reference, and a mathematical solution(s) presented. The project has been failing most recently due to insufficiencies in all auto-tracking features attempted, but this has been greatly simplified now to just a manual tracking system initially. It broadly consists of the following:
1) Inputting a video that could be as long a 10-15 seconds long;
2) Reformatting the video to add frame numbers and/or time code, and converting different input formats into a single workable format as warranted;
3) An ability to analyze the reformatted video. Efficient operational speed and frame-by-frame viewing both forward and backward is crucial, with a feature to download the reformatted video (if a web application) for analysis in a local player if desired;
4) A section to input premeasured data about the club analyzed, including but not limited to club locations where a user attaches tracking markers to reference, and a connected interactive graphic that simulates the club and displays the information input;
5) An ability to select desired frames/times from the reformatted video for further analysis in the application. Selection could be a range of frames or any number of frames individually selected as examples;
6) An ability to view and manually, efficiently place and adjust club tracking points over the reference tracking markers in each of the selected frames. No auto-tracking system is needed as the number of frames to be analyzed will initially be quite limited;
7) Creating an interactive graphic of club movement based upon the tracking points set in 6), and;
8) Using the tracking data to calculate and display rotation point location values along the length of the club for club movement.
Substantial additional documentation is available for interested and qualified parties. With numerous past failures, any "official" budget was exhausted long ago and so funds are extremely limited.
I am open to a web or desktop version, with initial preferences toward which could be developed more quickly and less expensively and/or which could be studied for proper operation and improvements more conveniently.
An NDA would be appreciated before additional documentation is provided. Hourly or project quotes can be considered. Please feel free to submit any questions and/or comments.
Thank you very much.