Shivaram Venkataraman
Shivaram Venkataraman
Right now it assumes small image sizes and doesn't work when the number of pixels is > blockSize cc @ericmjonas
Based on discussion in #72 we should augment our SIFT, FV test cases to check binary equivalence with some known golden test cases
As discussed in #80 There are a couple of issues with running a pipeline on variable sized images 1. The convolver right now assumes images are of same size and...
Right now we only extract LCS features at a single scale. We should either use image resize or investigate other techniques to get LCS at more scales