Hi Sir,
Below is the question
You decide to use data augmentation to address foggy images. You find 1,000 pictures of fog off the internet, and “add” them to clean images to synthesize foggy days, like this:
To the above question, this option also reasonably correct right next to human eye realistic, because we have only one road image so adding 10,000 different fog image to same road image , the algorithm overfit right ?