Object detection in video

[chantzeleong]

How would the code be for object detection in video files? Any preferred video format for better accuracy?

[glenn-jocher]

@chantzeleong you might want to read the README:
https://github.com/ultralytics/yolov5#inference

Code is:

python detect.py --source video.mp4

[muhammadumair2019]

@MGIbrahim how did you do it then? video inference by loading model with pytorch not giving results. What should i do?

[MGIbrahim]

@muhammadumair2019 I had to divide the video into frames and pass it into the trained model using pytorch. I then would re-construct the video back from the frames after passing it to model.

[muhammadumair2019]

@MGIbrahim Thanks for your reply. But, I have to give a live video to the model for detection. In this case, it is a long process.
Do you know any faster way?

[MGIbrahim]

Maybe this could help? #2045

[muhammadumair2019]

Thankyou brother. It is very helpful.