First off, excellent project and a very detailed and well explained code.
I executed the code and trained it but ended up with a few incorrect recognition results. So, I tried to test with the pretrained models first.
What I did was just run inference_with_net.py with the pre trained ssd and rcnn models on two of the bmp images 14 and 36. It does not seem to be predicting the correct bounding box.
I want to know whether I have done something wrong or missed something?
Hi,
First off, excellent project and a very detailed and well explained code.
I executed the code and trained it but ended up with a few incorrect recognition results. So, I tried to test with the pretrained models first.
What I did was just run inference_with_net.py with the pre trained ssd and rcnn models on two of the bmp images 14 and 36. It does not seem to be predicting the correct bounding box.
I want to know whether I have done something wrong or missed something?