Event Image Retrieval using Decision Trees

Abstract: Efficient Image data storage and tagged Image Archive are quintessential in organizations due to the growing volume of images that are captured during various events that are conducted in organizations. Acquired Images could be indexed instantly using time stamp, image content and image context by means of object identification and association, face detection and matching, object feature extraction and identification. The indexing done using the foresaid criteria transforms each image into a feature vector with domain dependent attributes and is stored in a relational database with the image URL to felicitate query based image retrieval. The domain specific user query is designed to enable the query engine to extract each key attribute as per the weight associated in the proposed model. The efficiency and relevance of an image retrieval system rely on the visualization and the presentation of the search results. Thus the query matching with the feature vector is implemented by building a hierarchical decision tree. Keywords: CBIR, SIFT, QBIR, HOG, ROI, SSIM.