After Data Explorer visualizes the dataset, it provides us with the option of searching for specific examples.
In this case, we’re looking at a dataset of different surfaces, and this image has a horizontal fracture. We can mark it and search based on this image alone.
Naturally, searching based on one image might not provide the best examples, so Data Explorer provides us with the option of refining the search based on positive reinforcement as well as negative reinforcement.
We can refine the search further until we get the desired subset, and later we can export it and proceed as desired.
Here is an example of what we got based on just three images, and naturally, we can continue to refine.