I have extracted features such as sift,surf,edge,color and texture;and then I made a hash code using MD5. I have tried to compare and rank the bulk of images with query image with binary level hamming distance of feature hash code and direct hash code hamming distance with each feature in each time.Unfortunately these two hamming distances are giving me worst result in the case of every feature hash codes.
I just wanna solution for hash code based image search in java.And want to know how can i generate bitwise weight for each hash code.