Hello guys, does anyone of you know how to increase the dataset by using image augmentation?
@Irsyad21 you can use javacv for that which we already depend on in dl4j. The idea with image augmentation is you modify the image using something in opencv and then save it as a new example.
Usually with javacv that would be imwrite. The types of augmentation you should do would depend on your application. Here’s an example in python that maps pretty much 1 to 1 with javacv if you want to try it: