Face Image Retrieval Based on Fireworks Algorithm

Content-Based Image Retrieval (CBIR) is a technique used in search engine like google, yahoo, etc. for retrieving most relevant images depending on feeding information that may be text or image. The classical methods of image indexing in the database used keywords or numbers with description record, but these methods have become outmoded. This paper deals with retrieving face images from the datacenter that related to the input image. The proposed method used object detection to specify the object in the image (human face) using the viola-jones algorithm. Local Binary Pattern (LBP) is used for feature extraction and store the features in the database. The used method for image features clustering is K-means method. The proposed system optimizes the total features of each cluster using Firework algorithm (FWA) that work on reducing the search space and constructing a temporal table utilized as intermediate indexing of related image. The services provided by this system that are loading the related databases only in memory for query image and retrieve the most relevant in sorted order depend on distance measurement. The obtained results indicated that the proposed system provides an efficient detection for a person in the whole data center if there is one person available in the image or more than one person related to searching image (query image).