Object-Detection

Real-time Object Detection using state-of-the-art You Only Look Once (YOLO) algorithm.

Requirements

1. Tesorflow/Tensorflow-gpu (Compatibility)
2. openCV
3. Setting up Darkflow
4. Weights and cfg file

• Create a weights folder in darkflow and paste the weights in there.

Processing a video using cmd.

• Move the video into darkflow-master.
• Open cmd from there and use the command.

python flow --model cfg/yolo.cfg --load weights/yolov2.weights --demo "video.mp4" --gpu 0.6 --saveVideo

• Enter your video file name instead of "video.mp4"
• Leave off --gpu 0.6 if you are not using Tensorflow-gpu.

Swap out real-time detection to process a video.

capture = cv2.VideoCapture(filename)