Python Image Processing Cookbook

Implementing an image search engine

In this recipe, you will learn how to implement a simple image search engine (SE). It will be a search by example system, relying only on the image contents, known as content-based image retrieval (CBIR) systems. The images along with the features extracted are stored so that the system can return similar images (based on the features) during a search. The following describes the four steps of any CBIR system:

1. Defining an image descriptor (descriptive features of an image)
2. Indexing search images (for quick retrieval of the images with similar descriptors to the query image. Use an efficient data structure for fast retrieval)
3. Defining the similarity metric to be used (Euclidean/cosine/chi-squared distance, ...