Object detection web-page based on flask, opencv and tiny YOLO
Convert weights
copy weights and names files into ./tensorflow-yolo-v3/model_files
cd tensorflow-yolo-v3 &&
python3 convert_weights_pb.py \
--class_names ./model_files/obj.names \
--weights_file ./model_files/yolov3-tiny-face-palms.weights \
--data_format NHWC \
--tiny
Run
python app.py
Acknowledgments
The simplest implementation of the flask webcam streamer: streaming_webcam_with_flask
YOLOv3 weights converter: tensorflow-yolo-v3
Custom TensorFlow binaries for old CPU without AVX instruction suppot: TensorFlow binaries supporting AVX, FMA, SSE
TensorFlow YOLOV3/TinyYOLOV3 implementation with tutirials: YOLOv3_TF2