I need a software that runs on AWS (amazon web service) and recognizes structures, people, OCR text from video files as accurately as possible.
Input: video files in any format
Output: metadata that provides the following
[time frame within video] [video frame identification] [x,y coordinates] [recognized data in that x,y coordinate]
You can use OpenCV for object recognition or any other image recognition tools out there.
Please contact me for more details on the project