3 ms·
In theory, yes, you can feed these new distorted images into a new model and defeat the cloaking. In practice, building new training sets is much more expensiv
by scribu 6y ago
In theory, yes, you can feed these new distorted images into a new model and defeat the cloaking.
In practice, building new training sets is much more expensive than slightly changing the cloaking algorithm.