Content Based Image & Video Retrieval
Now a day’s digital data like images and videos increasing too fast which causes much research effort to be
devoted to develop video and image retrieval methods to retrieve video or image of interest. Traditionally text-based
retrieval system is used to retrieve video or images from database but this is not efficient approach so to address this
problems associated with traditional system Content Based Image Retrieval (CBIR) and Content Based Video Retrieval
(CBVR) were introduced. In several applications of CBIR & CBIR system has been used, such as crime prevention,
fingerprint identification, digital libraries, medicine, historical research, video on demand service, etc. Basically, CBIR &
CBVR system tries to retrieve images or videos similar to a user-query and their goal is to retrieve similar image or video
based on content properties such as shape, colour, texture, motion. Content properties typically set into the feature vectors.
In this dissertation, plan to implement, Methodology for CBIR based on image classification using Support Vector Machine
(SVM) classifier is introduced and CBIR used C4.5 classifier. Main purpose of this approach is to narrow down the search
space. HSV technique for colour feature, Gabion filter technique for texture detection, Eccentricity for shape, motion
detection we will use background subtraction technique, Kd- tree for indexing.
Keywords - CBIR, CBVR, Gabor filter, Color moment, Eccentricity, SVM, Colour Feature Extraction, Motion Feature,
Extraction, Shape Feature Extraction, Texture Feature.