Yes, detection is performed on every frame read by "cap.read()". Detecting on every frame can be computationally intensive for longer or high-resolution videos. You may optimize this by skipping frames or resizing frames for faster processing if real-time speed isn’t critical. Since i am using "model.track()", the detections in consecutive frames are linked to enable consistent object identification across the video
nice. interesting to see all code + make some experiments and integration
Thank !
You can se all code on my GitHub/glennraana
did you detect in each frame
Yes, detection is performed on every frame read by "cap.read()".
Detecting on every frame can be computationally intensive for longer or high-resolution videos. You may optimize this by skipping frames or resizing frames for faster processing if real-time speed isn’t critical.
Since i am using "model.track()", the detections in consecutive frames are linked to enable consistent object identification across the video
Can i have your full code?
Hi :) all the code for projects on youtube is here: github.com/glennraana