IVSS: Integration of Color Feature Extraction Techniques for Intelligent Video Search Systems Avinash N Bhute Dept. of Computer Technology Veermata Jijabai Technological Institute Matunga, Mumbai, India e-mail: anbhute@gmail.com B. B. Meshram Dept. of Computer Technology Veermata Jijabai Technological Institute Matunga, Mumbai, India e-mail: bbmeshram@vjti.org.in Abstract—As large amount of visual Information is available on web in form of images, graphics, animations and videos, so it is important in internet era to have an effective video search system. As there are number of video search engine (blinkx, Videosurf, Google, YouTube, etc.) which search for relevant videos based on user “keyword” or “term”, But very less commercial video search engine are available which search videos based on visual image/clip/video. In this paper we are recommending a system that will search for relevant video using color feature of video in response of user Query. Keywords- CBVR, Video Segmentation, Key Feature Extractio, Color Feature Extraction, Classification. I. INTRODUCTION As in internet era most difficult task is to retrieve the relevant information in response to a query. To help a user in this context various search system/engine are there in market with different features. In web search era 1.0 the main focus was on text retrieval using link analysis. It was totally read only era. There was no interaction in between the user and the search engine i.e. after obtaining search result user have no option to provide feedback regarding whether the result is relevant or not. In web search era 2.0 the focus was on retrieval of data based on relevance ranking as well as on social networking to read, write, edit and publish the result. Due to Proliferation of technology the current search era based on contextual search. Where rather than ranking of a page focus is on content based similarity to provide accurate result to user. The CBVR (Content Based Video Retrieval) have received intensive attention in the literature of video information retrieval since this area was started couple of years ago, and consequently a broad range of techniques has been proposed. The algorithms used in these systems are commonly divided into four tasks: • Segmentation • Extraction • Selection, and • Classification The segmentation task splits the video into number of chunks or shots. The extraction task transforms the content of video into various content features. Feature extraction is the process of generating features to be used in the selection and classification tasks. A feature is a characteristic that can capture a certain visual property of an image either globally for the whole image, or locally for objects or regions. Feature selection reduces the number of features provided to the classification task. Those features which are assisting in discrimination are selected and which are not selected is discarded. The selected features are used in the classification task [2]. The figure 1 shows the content based video search systems with four primitive tasks. Fig. 1 Content based Video Search Systems and Its Task Among these four activities feature extraction is critical because it directly influence the classification task. The set of features are the result of feature extraction. In past few years, the number of CBVR systems using different segmentation and extraction techniques, which proves reliable professional applications in Industry automation, social security, crime prevention, biometric security, CCTV surveillance [1], etc. II. PROPOSED CBVR SYSTEMS Due to rapidity of digital information (Audio, Video) it become essential to develop a tool for efficient search of these media. With help of this paper we are proposing a Video Search system which will provide accurate and efficient result to a user query. The proposed system is a web based application as shown in fig.1 which consists of following processing: 1) Client Side Processing: From client machine user can access the Graphical User Interface of the system. User can access and able to perform three tasks: a) Register the video b) Search the video c) Accept the efficient result from server. 2. Server Side Processing: The core processing will be carried out at server side to minimize the overhead on client. Client will make a request for similar type of videos by providing query by video clip. On reception of this query by video clip, server will perform some processing on query video as well as on videos in its database and extract the video which are similar to query video. After retrieving the similar videos from the database,