scalable-image-matching

This is a image matching system (library) for scalable and efficient matching of images from a large database. The basic idea is to compute perceptural hash value for each image and compare the similarity based on the pHash computed. Searching are scalable by using elasticsearch as the backend database.

Online search and local search are both supported, i.e. providing the URL or the local address of the target image to the library.

Get Started

• Clone this repo and enter the directory;
• Install with pip: pip install numpy scipy .;
• Install Elasticsearch, https://www.elastic.co/downloads/elasticsearch;
• Make sure elasticsearch is running before scalable searching is used;

Useful Links

Once you're up and running, read these two (short) sections of the documentation to get a feeling of what scalable-image-match is capable of:

• Image signatures
• Storing and searching images