Forked from qqwweee/keras-yolo3 to edit the training process and train with in order to detect the food
- Download the weight and stored in the /model_data/ and named it 'yolo_food.h5' Just run it using python yolo_video.py --image and input the image path
- The result can be with the file called detection.png
A UECFood256 Dataset
- Try to convert to TF-Serving format to deploy with Flask
- Upgrade the output to return the Bounding Box for all detection result
- Change from detect only one image to detect whole directory