Content Based Video Retrieval (CBVR)

especially images, music, and video, is quickly gaining importance for the business and entertainment industry. Content-based video retrieval (CBVR) is a prominent research interest. The ability of a computer to automatically recognize objects in videos is so low that the existing technique for extracting semantic features from all kinds of videos is incapable of retrieving videos based on semantic feature. Here a new approach is proposed for video clip retrieval based on Earth Mover’s Distance (EMD). Instead of imposing one-to-one matching constraint it allows many-to-many matching methodology and is capable of tolerating errors due to video partitioning and various video editing effects. We formulate clip-based retrieval as a graph matching problem in two stages. In the first stage, to allow the matching between a query and a long video, an online clip segmentation algorithm is employed to rapidly locate candidate clips for similarity measure. In the second stage, a weighted graph is constructed to model the similarity between two clips. EMD is proposed to compute the minimum cost of the weighted graph as the similarity between two clips.

Explosive growth of digital content including image, audio and video on web and as well as on desktop applications has demanded development of new technologies and methods for representation, storage and retrieval of multimedia data. Rapid development of digital libraries and repositories are attempting to achieve efficient techniques for the same. Despite of many initial successes problems persist in the area of effective video retrieval system since decades. Many of the video retrieval systems are presently based on the metadata attributes like name, date of creation, tagged words, annotation, etc. This however leads to unsatisfactory results to users. Content Based Video Retrieval (CBVR) system works more effectively as these deals with content of video rather than video metadata. The increased availability and usage of on-line digital video has created a need for automated video content analysis techniques including indexing and retrieving.

By definition, a Content-Based Video Retrieval (CBVR) system aims at assisting a human operator to retrieve sequence (target) within a potentially large database [11]. The authors in [11] has just presented a natural extension of the well-known Content-Based Image Indexing and Retrieval (CBIR) systems. Both systems are aiming at accessing image and video by its content, namely, the spatial (image) and spatial-temporal (video) information. A Typical spatial information includes texture, color, edge, etc., while a typical temporal information includes change of scenes and motions. Moving from images to video adds several orders of complexity to the retrieval problem due to indexing, analysis and browsing over the inherently temporal aspect of video [15].

Content Based Video Retrieval (CBVR) has been increasingly used to describe the process of retrieving desired videos from a large collection on the basis of features that are extracted from the videos. The extracted features are used to index, classify and retrieve desired and relevant videos while filtering out undesired ones. Videos can be represented by their audio, texts, faces and objects in their frames. An individual video possesses unique motion features, color histograms, motion histograms, text features, audio features, features extracted from faces and objects existing in its frames. Videos containing useful information and occupying significant space in the databases are under-utilized unless CBVR systems capable of retrieving desired videos by sharply selecting relevant while filtering out undesired videos exist. Results have shown performance improvement (higher precision and recall values) when features suitable to particular types of videos are utilized wisely. Various combinations of these features can also be used to achieve desired performance. In this paper a complex and wide area of CBVR and CBVR systems has been presented in a comprehensive and simple way. Processes at different stages in CBVR systems are described in a systematic way. Types of features, their combinations and their utilization methods, techniques and algorithms are also shown. Various querying methods, some of the features like GLCM, Gabor Magnitude, algorithm to obtain similarity like Kullback-Leibler distance method and Relevance Feedback Method are discussed.

Video retrieval has recently attracted a lot of research attention due to the exponential growth of video datasets and the internet. Content based video retrieval (CBVR) systems are very useful for a wide range of applications with several type of data such as visual, audio and metadata. In this paper, we are only using the visual information from the video. Shot boundary detection, key frame extraction, and video retrieval are three important parts of CBVR systems. In this paper, we have modified and proposed new methods for the three important parts of our CBVR system. Meanwhile, the local and global color, texture, and motion features of the video are extracted as features of key frames. To evaluate the applicability of the proposed technique against various methods, the P(1) metric and the CC_WEB_VIDEO dataset are used. The experimental results show that the proposed method provides better performance and less processing time compared to the other methods.

Video has become an important element of multimedia computing and communication environments due to cheap devices like digital cameras, smart phones, etc. Due to these advances in transmission technologies, we are seeing the abrupt growth of videos in the social networking sites with less or without semantic tags associated with it. According to YouTube statistics, every minute about 200 hours of video content is being uploaded to a website like YouTube and similarly around 11 million videos are posted daily in twitter without texts or with poor semantic tags. Because of this explosive growth of online videos without semantic tags there is need of content based video retrieval (CBVR) in large scale. CBVR is problem of retrieving most similar videos to a given query video or image. ‘Content-based’ means that search is done by analyzing the visual content rather than metadata (like tags or description) associated with it. Here, the term ‘Content’ means visual features extracted

Abstract:- Traditional video retrieval methods fail to meet technical challenges due to large and rapid growth of multimedia data, demanding effective retrieval systems. In the last decade Content Based Video Retrieval (CBVR) has become more and more popular. The amount of lecture video data on the Worldwide Web (WWW) is growing rapidly. Therefore, a more efficient method for video retrieval in WWW or within large lecture video archives is urgently needed. This paper presents an implementation of automated video indexing and video search in large videodatabase. First of all, we apply automatic video segmentation and key-frame detection to extract the frames from video. At next, we extract textual keywords by applying on video i.e. Optical Character Recognition (OCR) technology on key-frames and Automatic Speech Recognition (ASR) on audio tracks of that video. At next, we also extractingcolour, texture and edge detector features from different method. At last, we integrate all the keywords and features which has extracted from above techniques for searching purpose.Finallysearch similarity measure is applied to retrieve the best matchingcorresponding videos are presented as output from database. Additionally we are providing Re-ranking of results as per users interest in original result.

object clusters. Object Based Video Retrieval is introduced in[11]. In this work Edge Detection and DCT based block matching is used for shot segmentation and the region based approach is used for retrieval. In content based Video Retrieval (CBVR) the feature extraction plays the main role. The features are extracted from the regions by using SIFT features. Finally the features of the query object are compared with the shot features for retrieval. Evaluation of Object Based Video Retrieval Using SIFT in introduced in [12]. The local invariant features are obtained for all frames in a sequence and tracked throughout the shot to extract stable features. Proposed work is to retrieve video from the database by giving query as an object. Video is firstly converted into frames, these frames are then segmented and an object is separated from the image. Then features are extracted from object image by using SIFT features. Features of the video database obtained by the segmentation and feature extraction using SIFT feature are matched by Nearest Neighbour Search (NNS). Spatiotemporal Region Graph Indexing for Large Video Databases is introduced in[13]. Here the authors propose a new graph-based data structure and indexing to organize and retrieve video data. Several re- searches have shown that a graph can be a better candidate for modelling semantically rich and complicated multimedia data. Proposed system uses a new graph- based data structure called Spatio-Temporal Region Graph (STRG). STRG further provides temporal features, which represent temporal relationships among spatial objects. The STRG is decomposed into its sub graphs in which redundant sub graphs are eliminated to reduce the index size and search time, because the computational complexity of graph matching (sub graph isomorphism) is NP-complete. In addition, a new distance measure, called Extended Graph Edit Distance (EGED), is introduced in both non-metric and metric spaces for matching and indexing respectively.

Content material-based retrieval lets in finding data with the aid of searching its content as opposed to its attributes. The undertaking dealing with content-based video retrieval (CBVR) is to layout systems that can accurately and routinely method huge amounts of heterogeneous motion pictures. Furthermore, content material-based video retrieval machine calls for in its first level to phase the video movement into separate shots. Afterwards functions are extracted for video pictures representation. And sooner or later, pick out a similarity/distance metric and an set of rules this is green sufficient to retrieve query – related videos effects. There are major problems in this manner; the primary is the way to decide the first-rate way for video segmentation and key body selection. The 2nd is the capabilities used for video illustration. Diverse features can be extracted for this sake which includes either low or high stage functions. A key problem is how to bridge the space between low and high level features. This paper proposes a gadget for a content based totally video retrieval system that tries to address the aforementioned troubles through the usage of adaptive threshold for video segmentation and key frame selection in addition to the usage of each low level features collectively with excessive degree semantic item annotation for video illustration. Experimental outcomes show that the use of multi features increases each precision and bear in mind rates via about 13% to 19 % than traditional gadget that uses best shade function for video retrieval.

content of any media is unachievable [2]. For this reason, we require a good search technique for Content-Based Video Retrieval System (CBVR). In other words, content-based is defined as the search which will examine the original image contents. Here, content relates to colors, shapes, textures, or any other information that can be obtained from the image directly [5]. Recently, CBVR system has been widely studied. In CBVR, vital information is automatically taken out by employing signal processing and pattern recognition techniques for audio and video signals [2]. Digital video needs to efficiently store the index, store, and retrieve the visual information from multimedia database. Video has both spatial and temporal dimensions and video index should capture the spatio-temporal contents of the scene. In order to achieve this, a framework mainly works into three basic steps. Shot segmentation, Feature extraction and finally similarity match for effective retrieval of the query clip. This approach has established a general framework of image retrieval from a new perspective. The query example may be an image, a shot or a clip. A shot is a sequence of frames that was continuously captured by the same camera, while a clip is a series of shots describing a particular event. Our query statement is formulated as: given a sample clip, find all occurrences of similar (or relevant) video clips in the database. Current techniques for content-based video retrieval can be broadly classified into two categories.

The multimedia storage grows and the cost for storing multimedia data is cheaper. So there is huge number of videos available in the video repositories. It is difficult to retrieve the relevant videos from large video repository as per user in-terest as users shift from text based retrieval systems to content based retrieval systems. Video retrieval is very important in multimedia database management. This paper offers an overview of the landscape of general strategies in visual content-based video retrieval, focusing on methods for video structure analysis, including shot boundary detection, key frame extraction, extraction of features including static key frame features, object feature, video retrieval including similarity measures and the proposed procedure consists of the unique aspect of clustering techniques.

Videos are a powerful and communicative media that can capture and present information. The rapidly expanding digital video information has motivated growth of new technologies for effective browsing, annotating and retrieval of video data. Content-based video retrieval has attracted wide research during the last 10 years. Users are more diverted to content based search rather than text based search. These lead to the process of selecting, indexing and ranking the database according to the human visual perception. This paper reviews the recent research in content based video retrieval system. This survey focusing on video structure analysis, like, shot boundary detection and key frame extraction, different feature extraction methods including SIFT, SURF, etc, similarity measure, video indexing, and video browsing. This system retrieves similar videos based on local feature descriptor called SURF (Speeded-Up Robust Feature). For image convolution SURF relies on integral images. In SURF we use Hessian matrix-based measure for the detector and a distribution-based descriptor. SURF can be computed and compared much faster with respect to repeatability, uniqueness and robustness. SURF is better than previous proposed methods as SIFT, PCA-SIFT, GLOH, etc. Finally the future scope in this system is specified.

Video browsing can be achieved by segmenting video into representative key frames. The selected key frames can provide a visual guideline for navigation in the lecture video portal. Moreover, video segmentation and key- frame selection is also often adopted as a pre-processing for other analysis tasks such as video OCR, visual concept detection, etc. Choosing a sufficient segmentation method is based on the definition of “video segment” and usually depends on the genre of the video. In the lecture video domain, the video sequence of an individual lecture topic or subtopic is often considered as a video segment In the first step, the entire slide video is analyzed. We try to capture every knowledge change between adjacent frames, for which we established an analysis interval of three seconds by taking both accuracy and efficiency into account. In the second segmentation step the real slide transitions will be captured. The title and content region of a slide frame is first defined.

Abstract: Key frame extraction has been recognized as one of the important research issues in video information retrieval. Although progress has been made in key frame extraction, the existing approaches are either computationally expensive or ineffective in capturing most important visual content. Video summarization aimed at reducing the amount of data that must be examined in order to retrieve the information desired from information in a video, is an essential task in video analysis and indexing applications. We propose an innovative approach to the selection of representative (key) frames of a video sequence for video summarization In this paper, we discuss the importance of key frame selection; and then briefly review and evaluate the existing approaches. To overcome the shortcomings of the existing approaches, we introduce a new algorithm for key frame extraction.

We propose an efficient CBVR (Content based video retrieval), for identifying and retrieving similar videos from very large video database.. Here searching is based on[r]

Content-based video retrieval (CBVR) is the application of computer vision techniques to the video footage image retrieval problem i.e. searching for digital video content in long time recording of a spot using close circuit TV camera e.g. in ATMs, shopping malls etc. Content-based method means that the search analyzes the contents of the video recording rather than the metadata search keywords, tags, or descriptions associated with the recording. The term "content" in this context refers to any unusual activity in an event for long time, like staying of a person for a long time or more than expected time of a person in an ATM.

Content-Based Video Retrieval (CBVR) has been increasingly used to describe the process of retrieving desired videos from a large collection on the basis of features that are extracted from the videos. The extracted features are used to index, classify and retrieve desired and relevant videos while filtering out undesired ones. Videos can be represented by their audio, texts, faces and objects in their frames. An individual video possesses unique motion features, color histograms, motion histograms, text features, audio features, features extracted from faces and objects existing in its frames. Videos containing useful information and occupying significant space in the databases are under-utilized unless CBVR systems capable of retrieving desired videos by sharply selecting relevant while filtering out undesired videos exist. Results have shown performance improvement (higher precision and recall values) when features suitable to particular types of videos are utilized wisely. Various combinations of these features can also be used to achieve desired performance. In this paper, a complex and wide area of CBVR and CBVR systems have been presented in a comprehensive and simple way. Processes at different stages in CBVR systems are described in a systematic way. Types of features, their combinations and their utilization methods, techniques and algorithms are also shown. Various querying methods, some of the features like GLCM, Gabor Magnitude, an algorithm to obtain similarities like Kullback-Leibler distance method and Relevance Feedback Method are discussed.

DOI: 10.4236/jcc.2018.68003 29 Journal of Computer and Communications Video retrieval is still an active problem due to the semantic gap, and the widespread of social media and the enormous technological development. Pro- viding an efficient video retrieval with these huge amounts of videos on the web or even stored on the storage media is a difficult problem. The causation of the semantic gap is the difference between user requirements which are represented in queries and the low-level representation of videos on the storage media. Many methods are proposed to solve this semantic gap [2]-[7], etc., but it is not fully bridged. In this paper, a concise overview of the content-based video retrieval is mentioned. After that, the definition and the causes of a semantic gap in video retrieval will be explored. As the concept detectors [8] play a vital role in seman- tic video retrieval, a thorough study of the obstacles that face the construction of the generic concept detectors will be presented. Finally, the different methods model semantic concept relationships in video retrieval are categorized and ex- plained in more details.

sequence which is required in some real life applications. In the text-based video retrieval system, videos can be retrieved based on keywords, i.e., video name as an input and based on the name, videos having similar names are retrieved. For example, to search a video of a particular cricket match from a large database, suppose the input as 'ICC world cup final cricket match'. But, if the videos in the database have same string as tag or keyword, system will display many results which may contain many irrelevant videos. If the keyword for a tennis match video is assigned as 'cricket match', the irrelevant video of tennis match is retrieved when user searching for 'cricket match'. To improve the efficiency of these systems, ‘Content-Based Video Retrieval Systems’[2] are developed. This project involves implementation of a Content-Based Video Retrieval Systems using video similarity algorithm which provides the similarity measure between two video sequences.

User gets frustrated with duplicate contents. Universally consented video retrieval and indexing methods are not well defined or available. Hence a content based video retrieval system is presented as a motivation from above challenges. Content Based Video Retrieval system it includes various steps like Video Segmentation, Key frame Selection, Feature Extraction, Classification and Clustering, Indexing and Similarity.