This code base is yolov5 but i just add some parts to adding tranparet image to our train data for increasing our MAP. We start to reconizge birds in cage but our main porblem was about our data set.
That was really hard to find a data base for learning this kind of model.
So, we just follow our simple idea to add cage to data set which had normal birds image in it.
We have 2 kinds of way for doing this idea:
1.Create data set with image that we add them cages and then train it.
2.Add cage images in Load_Image function in yolov5 during trainig.
All of the libererys are exatlly same as Yolov5 :
git clone https://github.com/ShAmoNiA/yolov5_addingCage
cd yolov5_addingCage
pip install -r requirements.txt
For running GA alghorithm: