The images captured previously are slightly dark. If we increase the brightness during actual operation, will it affect the results of the deep learning model? Do we need to retrain the deep learning model?
Yes, the deep learning model is trained using 2D images. When the lighting varies, the features under different lighting conditions may change. However, the model will still rely on its previous feature memory to detect edges. In cases where it can’t identify them, it might result in the loss of certain features.