tejaslodaya/car-detection-yolo

Autonomous driving - car detection using the very powerful YOLO model

car-detection-yolo's Issues

yolo.h5

Hey, I really liked your work. I am relatively new to DL and Transfer learning and Keras, could you guide me as to how do I generate yolo.h5 or share an equally convenient file with me given the urgency I'm in? I did follow the link the lead me to a whole new project about which I have no idea.
Thanks in advance.

may i get onnx model?

hi , i am really interested in your project , may i get this model onnx format?

'list' object has no attribute 'dtype' @ yolo_head(yolo_model.output, anchors, len(class_names))

yolo_head(yolo_model.output, anchors, len(class_names)),
gives an error 'list' object has no attribute 'dtype'.
Issue was closed according to qqwweee/keras-yolo3#336 , but the remedy seems not implemented over here

Fix inter_area

car-detection-yolo/app_utils.py

Line 56 in 3d32c33

inter_area = (yi2 - yi1) * (xi2 - xi1)

There should be max((yi2 - yi1) , 0) * max((xi2 - xi1), 0), take into account that this value MUST be positive or 0

regarding non-max suppression in README

Hi,

Thank you for the writeup ! It's really useful.
I have one question regarding non-max suppression interpretation you wrote in README :

You mentioned..

1,Select the box that has the highest score.
2,Compute its overlap with all other boxes, and remove boxes that overlap it more than iou_threshold.
3,Iterate until there's no more boxes with a lower score than the current selected box.

My question is about step3.
What if there is a ,say, car in the up right corner of the pic with prob = 0.8, and another car in the up left corner of the pic with prob = 0.6. They are not overlapped. ( IOU = 0 )
We should keep both of them

But...if we do as what step3 says, the algorithm will remove the prob = 0.6 box right ?