A SURVEY ON CONTENT BASED VIDEO RETRIEVAL AND ANALYSIS
Shivappa M Metagar, Anil S. Naik, Vishwanath Chavan · 2016
Content based video retrieval and analysis is the one of the most important and recent research area in the Image processing domain. Content based video retrieval means whatever the input is given by the user based on that input, we will take video and analyze that video, actually what we have to analyze in that video is features of the video like texts, images, colours, content of the video and image movements in that video etc. Content based video retrieval can be used for multiuser systems for video search and browsing which are useful in web applications. Content based video retrieval methods are applied in order to improve the effectiveness of the content based methods. And these methods can play essential rules which are used in media collection and enhance retrieval accuracy. Similarly we will study the other topics like content based image retrieval and document based image retrieval to comparing these topics content based video retrieval is little bit difficult and complex topic. And Content based video retrieval may be defined as an approach in which the videos are retrieved from the large database based upon their visual contents. General terms: CBVR, video retrieval, shots, feature extraction. Developed by academic institutes such as Photobook, Netra,Visualseek and Chabot to check new technologies. In the context of multimedia, information retrieval such as images, text, audio or videos is an open research area. Various methods have been discovered to retrieve the effective and accurate information. Basically, there are two retrieval methods, which are: text based information retrieval and content based video retrieval. The First method is the text based retrieval- in which the images and videos are manually annotated with keywords or descriptors (1). Most of the search engines on the internet like Google tend to find the multimedia information by searching the textual labels attached with the multimedia data. The keywords are attached to the multimedia data in such a way that they tend to best describe the image or video itself based on their properties. Second method is content based information retrieval, which deals with retrieval of images or videos based on their visual contents from large database of videos. With the help of content based video retrieval, the user is able to retrieve important clips of video based on his demands rather than watching the whole video. So, content based video indexing and retrieval is promising area of research. There are three levels to describe visual features of image or video described below: Low level description Middle level description High level description Low level description of the visual contents is based on color, texture and shape features. The low level visual features of an image are directly related to the image content. Next level is the middle level description which is concerned with the background and spatial attributes of the concerned object. The high level feature representation is dealing with human brain and perception. Examples are events, scenes and human thinking such as emotions (3). This type of description is very difficult to map to some mathematical model. A lot of work has been done in the field of content based image and video retrieval systems. Some of the important commercial systems are