ONLY developers with REAL working experience in this may apply. This is an urgent project, so there is no time for experimentation.
- Detect objects and track the movement to 90% + accuracy
- Detected objects to be classified . for eg : Vehicles as Car, Motorbike, SUV, Truck, Bus etc and Humans by gender, Age group, Actions such as walking, running, sitting. and other stationary objects.
- To be processed in real time with a maximum of +1 second delay
- Should be able to take all video formats
- Should be able to process videos from all protocols RTSP, HTTP, HTTPS, Single video upload, read from Folder,Youtube videos, Youtube Live feeds
- The script to take input in the form of config file. The config file to have details of the input source, ojbjects to detect etc
- The script to produce CSV output with detected details
- Should be in Python and Linux code
- Code should be optimised to run 4 parallel video feeds with near real time results on a quad core CPU with 4GB Ram Graphics card.