Build Python program to generate output from high res source (video)
Budget: $750 – $1,500 AUD
Need a Python(3) program built, using opencv (and possibly Tiny Yolo/YoloV3),
to detect objects from source video that is 3840x1088 (panoramic), then output
an HD frame (1920x1080 or 1280x720) based on that object detection and tracking.
Alternative option of using Goggle Mediapipes Autoflip (may produce the result we need),
however the program would probably then need to be C+ based (autoflip support is only C+ current,
not Python or JS, etc, like some other Mediapipe solutions).
Output is required to be output as file or stream (gstreamer or ffmpeg output is fine).
This is also to be run on a NVidia Jetson Nano, so that must be a consideration in coding (with use
of Jetson multimedia API/Jetpack libs to ensure efficient use of system resources).
to detect objects from source video that is 3840x1088 (panoramic), then output
an HD frame (1920x1080 or 1280x720) based on that object detection and tracking.
Alternative option of using Goggle Mediapipes Autoflip (may produce the result we need),
however the program would probably then need to be C+ based (autoflip support is only C+ current,
not Python or JS, etc, like some other Mediapipe solutions).
Output is required to be output as file or stream (gstreamer or ffmpeg output is fine).
This is also to be run on a NVidia Jetson Nano, so that must be a consideration in coding (with use
of Jetson multimedia API/Jetpack libs to ensure efficient use of system resources).