This repository contains the implementation of a image-based search engine using the normalized Hamming distance of the Perception Hashes of images to perform KNN and Range Search.
The data structure used to store the pHashes is the Vantage-Point Tree.
An example for serving the search engine can be found inside src/example. The
application will load a tab separated file called load_file_phash.csv (not provided) containing
two columns: path and phash, where the path refers to the path of the image, and the phash
refers to its Perception Hash. It also provides a File System to store uploaded images under the
relative directory images/temp/.
To begin serving the example engine,
cd src/example
go run .
A simple UI is implemented in React JS under app/. The UI lets you upload an image, and search
similar image against the engine using KNN or Range Search. To start the React app,
cd app
npm start