Edit Question - KnowledgeHub

Title Body

its about the libpuzzle libray for php ( http://libpuzzle.pureftpd.org/project/libpuzzle ) from Mr. Frank Denis. I´am trying to understand how to index and store the data in my mysql database. The generation of the vector is absolutly no problem. Example: # Compute signatures for two images $cvec1 = puzzle_fill_cvec_from_file('img1.jpg'); $cvec2 = puzzle_fill_cvec_from_file('img2.jpg'); # Compute the distance between both signatures $d = puzzle_vector_normalized_distance($cvec1, $cvec2); # Are pictures similar? if ($d < PUZZLE_CVEC_SIMILARITY_LOWER_THRESHOLD) { echo "Pictures are looking similar\n"; } else { echo "Pictures are different, distance=$d\n"; } Thats all clear to me - but now how do i work when i have a big amount of pictures >1.000.000? I calculate the vector and store it with the filename in the database? How to find the similar pictures now? If i store every vector in the mysql i have to open each record and calculate the distance with the puzzle_vector_normalized_distance function. That procedures takes alot of time (open each database entry - put it throw the function ,...) I read the readme from the lib puzzle libaray and found the following: Will it work with a database that has millions of pictures? A typical image signature only requires 182 bytes, using the built-in compression/decompression functions. Similar signatures share identical “words”, ie. identical sequences of values at the same positions. By using compound indexes (word + position), the set of possible similar vectors is dramatically reduced, and in most cases, no vector distance actually requires to get computed. Indexing through words and positions also makes it easy to split the data into multiple tables and servers. So yes, the Puzzle library is certainely not incompatible w

Tags (comma-separated) Cancel