You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Should be able to attempt to identify the line drawing and discard/send those to another algo. Then should be able to classify the rest of the images into buckets. Maybe while reducing the images to 3x 32x32 grids of rgb values, we could also acquire an "average" with std-dev or so and classify these images into buckets. Because there is no way a mostly white image is going to pass the similarity threshold of a almost black image. So this should be a good way to classify and bucket these images.
This way we can reduce the N of the comparing pool even further to make it even faster.
The text was updated successfully, but these errors were encountered:
This current SAD algo does not work with:
Should be able to attempt to identify the line drawing and discard/send those to another algo. Then should be able to classify the rest of the images into buckets. Maybe while reducing the images to 3x 32x32 grids of rgb values, we could also acquire an "average" with std-dev or so and classify these images into buckets. Because there is no way a mostly white image is going to pass the similarity threshold of a almost black image. So this should be a good way to classify and bucket these images.
This way we can reduce the N of the comparing pool even further to make it even faster.
The text was updated successfully, but these errors were encountered: