java - Using pHash to search agaist a huge image database, what is the best approach? -
i need search huge image database find possible duplicate using phash assuming image records have hash code generated using phash.
now have compare new image , have create hash using phash against existing records. per understanding has comparison not straight forward like
hash1 - has2 < threshold
looks need pass both hash codes phash api matching.so have retrieve hash codes db in batches , compare 1 one using phash api.
but looks not best approach if have 1000 images in queue compared against millions of exiting images.
i need know followings.
- is understanding/approach on using phash compare existing image db correct?
- is there better approach handle (without using cbir libraries lire)?
- i heard there algorithm called dhash can used image comparison hash codes..is there java libraries , can used phash optimize task of large image , repeated image processing tasks.
thanks in advance.
i think part of question discussed on phash support forum.
you need use mvptree storage mechanism
http://lists.phash.org/htdig.cgi/phash-support-phash.org/2011-may/000122.html , http://lists.phash.org/htdig.cgi/phash-support-phash.org/2010-october/000103.html
Comments
Post a Comment