Confusion matrix training statistics
I've been training this model on pothole images, but every time I train it, I always get 100% of the background images to be classified as pothole images.
Because the model works extremely well classifying potholes, why is the confusion matrix saying that all background images are classified as potholes? Also when I test it manually it seems to classify background images as background images and not give a bounding box to the pothole images. This is for pothole DETECTION.