emantic concept detection represents a key requirement in accessing large collections of digital mages/videos. Automatic detection of presence of a large number of semantic concepts, such as person,” or “waterfront,” or “explosion”, allows intuitive indexing and retrieval of visual content t the semantic level. Development of effective concept detectors and systematic evaluation ethods has become an active research topic in recent years. For example, a major video retrieval enchmarking event, NIST TRECVID[1], has contributed to this emerging area through (1) the rovision of large sets of common data and (2) the organization of common benchmark tasks to erform over this data.

With the growing popularity of video sharing websites and editing tools, it is easy for people to involve the video content from different sources into their own work, which raises the copyright problem. Contentbased video copy detection attempts to track the usage of the copyright-protected video content by using video analysis techniques, which deals with not only whether a copy occurs in a query video stream but also where the copy is located and where the copy is originated from. While a lot of work has addressed the problem with good performance, less effort has been made to consider the copy detection problem in the case of a continuous query stream, for which precise temporal localization and some complex video transformations like frame insertion and video editing need to be handled. In this chapter, the authors attack the problem by employing the graphical model to facilitate the frame fusion based video copy detection approach. The key idea is to convert frame fusion problem into graph model decoding problem with the temporal consistency constraint and three relaxed constraints. This work employs the HMM model to perform frame fusion and propose a Viterbi-like algorithm to speedup frame fusion process.

We submitted six interactive search runs for TRECVID2007, including 2 single user and 4 collaborative runs. In one single-user run (FXPAL MMA) the searchers had access to all resources including text search, text similarity, image similarity, and concept similarity search. In the other single-user submission (FXPAL MMV) the transcripts were not available and only features derived directly from the audio-visual content were available. This is a typical baseline for the manual and automatic search systems, but has been a less popular variation among the interactive search submissions. The other submitted runs (FXPAL CO*) employed a novel real-time, multi-user, collaborative search system. The complete set of submitted runs in priority order with system names and brief descriptions:

We present a new approach for video shot detection and introduce multi temporal distance images (MTDIs), formed by chisquare based similarity measures that are calculated pairwise within a floating window of video frames. By using MTDI-based boundary detectors, various cuts and transitions in various shapes (dissolves, overlayed effects, fades, and others) can be determined. The algorithm has been developed within the special context of soccer game TV broadcasts, where a particular interest in long view shots is intrinsic. With a correct shot detection rate in camera 1 shots of 98.2% within our representative test data set, our system outperforms competing state-of-the-art systems.

We present a general approach to temporal media segmentation using supervised classification. Given standard low-level features representing each time sample, we build intermediate features via pairwise similarity. The intermediate features comprehensively characterize local temporal structure, and are input to an efficient supervised classifier to identify shot boundaries. We integrate discriminative feature selection based on mutual information to enhance performance and reduce processing requirements. Experimental results using large-scale test sets provided by the TRECVID evaluations for abrupt and gradual shot boundary detection are presented, demonstrating excellent performance

We participated in the high-level feature extraction and search task for TRECVID 2005. For the high-level feature extraction task, we make use of the available collaborative annotation results for training, and develop 2 methods to perform automated concept annotation: (a) a ranked-Maximal Figure -of-Merit (MFoM) method; and (b) a multimodal rankBoost fusion method. We submitted a total of 7 runs based on these two methods. For the search task, we focus on improving our previous retrieval system by utilizing an event entity model derived from relevant external resources. In addition, we also make use of the various high-level feature extraction results contributed by various participating groups to help in the re-ranking step. We submitted a total 6 runs in the automated search category. The e valuation results show that our event-based approach is effective in human/event queries and that the high-level features is useful for general queries. 1. HIGH LEVEL FEATURE EXTRACTION TASK We explore two methods to perform high-level feature extraction. The first is based on a ranked-Maximal Figure-ofMerit (MFoM) method that has been successfully employed in text categorization. The second employs HMM for high-level features extraction, follow by rankBoost fusion to fuse with other modality features. 1.1 Ranked Maximal Figure -of-merit (MFoM) Conventional approaches for semantic concept detection is to train a binary classifier (e.g. SVM, Boosting) from the training set by optimizing generalized classification error or maximizing the likelihood. In the high-level extraction task, our concern is ranking the relevant shots as high as possible. Therefore, the basis of this approach lies in learning an optimal ranking function in terms of the mean average precision (MAP) from the development dataset, given any type of multimedia content representation (text from ASR, visual feature, etc). Here we develop an algorithm for training the ranking function with the goal of optimizing the MAP. This algorithm is similar to our work in (Gao et al., 2003 & 2004) and the ROC optimization for classifier (Cortes & Mohri, 2003; Yan, et al, 2003), where the objective function to be optimized is derived using an approximation of the interested metric for evaluation. A good measure for ranking is the Wilcoxon-Mann-Whitney statistic, which is equal to the area under the ROC (AUC), defined over the training set as: ( ) mn y x I U m i n j j i ∑ ∑ = = = 1 , (1) where, ( ) 1 : , 0: Otherwises i j i j x y I x y >  =   , with i x and j y being the scores from the classifier for the i-th out of the M positive samples, and the j-th out of the N negative examples. A classifier is then trained by maximizing Eq. (1) using a gradient algorithm. A sigmoid function (see Eq. (2)) is used to approximate the correct ranking count, ( ) j i y x I , , with ( ) ( ) j i y x j i e y x S − − + = β 1 1 , (2) where β is a constant function. After smoothing, Eq. (1) becomes a differentiable function. The smoothed Eq. (1) is a function of the parameters of the classifier (embedded by i x and j y ) and will be the objective for optimisation. Because it is highly non-linear, the gradient descent algorithm is applied to find its solution as in (Gao et al, 2003 & 2004). The ranking optimisation algorithm, n amed Rank-MFoM, is derived from MFoM learning in (Gao et al, 2003 & 2004). We submitted 4 runs using the Ranked-MFoM algorithm. They are based on: (a) only text feature; (b) only texture feature; and (c-d) fusions of both features using two different settings (e.g. β , the learning rate, and the iteration cycles) for the Rank-MFoM algorithm. The results are presented below. Run A (TRECVID Run 6): The text only run. The linear classifier is trained by Rank-MFoM on the shot-level text documents comprising the ASR outputs within 3-window shots. A lexicon with 3,464 terms is extracted from the ASR outputs in the development set after the removal of stop words and both very rare and frequent words. Each text document is represented using the tf-idf feature. In this run, shots without the ASR outputs are ignored. Run B (TRECVID Run 7): The texture only run. The image is uniformly segmented into 77 grids with a size 32x32. We extract a 12-dimensional texture feature (energy of log Gabor filter) from each grid. Subsequently, we apply Gaussian Mixture Model (GMM) to model the texture distributions for the positive and negative shots. The ranking function is the log-likelihood ratio between the positive and the negative class, defined for an image as, ( ) ( ) ( ) ( ) ∑ ∑ = = − = 771 1 771 0 log log 77 1 t t t t C x P C x P X L (3) with ( ) ( ) 1 , M i i i i j j j j p x C w μ = = ⋅ Ν Σ ∑ , 0,1 i = (4) where M is the number of Gaussian mixture (here we use M=4); ( ) , j j i i N μ Σ is the j-th Gaussian distribution with the mean, j i μ , and covariance, j i Σ , for the positive i-th class; and j i w the corresponding weight for each Gaussian component. Here C0 is the positive class and C1 is the negative class . For each concept, the parameters, j i w , j i μ and j i Σ , are estimated using the Rank-MFoM from images available in the development set. Run C (TRECVID Run 5): The fusion run 1. Since the output of the texture-based classifier is at the key-frame level, we simply choose the maximal likelihood ratio score among all key-frames for a shot as its shot-level output. We then combine the output of the text -based classifier to form a 2-dimensional vector. For shots without the ASR outputs, we treat it as the missing feature rather than removing them from the training set for fusion. The 2dimension vector is fused by training a single mixture GMM using the Rank-MFoM algorithm. Run D (TRECVID Run 3): The fusion run 2. This run is different from Run C as the shot-level score is taken to be the average of all key-frames for the shot. To illustrates the effectiveness of Rank-MFoM as compared to standard SVM using text feature, we perform a preliminary run on a subset of videos in the development set. Table 1 tabulates the MAP scores for the 10 concepts for the top 2000 shots. The results indicate that Rank-MFoM out-performs SVM on auto concept annotation task. Table 1: Preliminary results for Rank-MFoM vs. SVM on a subset of TRECVID 2005 Development Set 38 39 40 41 42 43 44 45 46 47 All Rank-MFoM 0.0091 0.0158 0.0759 0.0291 0.0076 0.0083 0.0175 0.0004 0.0997 0.0399 0.0303 SVM 0.0060 0.0102 0.0045 0.0003 0.0037 0.0091 0.0055 0.0000 0.0405 0.0133 0.0093 Table 2: Performance of Runs in terms of MAP from TRECVID evaluation 38 39 40 41 42 43 44 45 46 47 All Run A 0.0449 0.0169 0.0907 0.0498 0.0703 0.0231 0.0287 0.0054 0.0735 0.0638 0.0467 Run B 0.0727 0.0198 0.0527 0.0155 0.0990 0.1143 0.0693 0.0033 0.0769 0.0412 0.0565 Run C 0.0999 0.0426 0.1540 0.0809 0.1326 0.0989 0.0962 0.0102 0.1906 0.1189 0.1025 Run D 0.0872 0.0323 0.1575 0.0795 0.1332 0.0887 0.0939 0.0072 0.2249 0.1070 0.1011 Table 2 lists the results of the 4 runs on the TRECVID 2005 test set. The results clearly show that the fusion runs (Run C and Run D) perform significantly better than those runs employing only individual feature (Run A and Run B). The overall MAP of the texture-feature-only run is also slightly better than that of the text -feature-only run. We observe from the results that fusion has a positive effect on the overall performance, except for concept 43, i.e Waterscape/waterfront. The results show that the Rank-MFoM is effective in fusing different modality features for the detection problem. For these submissions, we only focus on using the extracted text and texture features. In our future work, more features will be introduced and fused. In addition, we will look at the association between different modalities and features to further improve the detection performance. 1.2 Multimodal RankBoost Fusion This approach aims to fuse a combination of both low and high level features using the rankBoost algorithm. The list of features to be fused include: text features from ASR, audio genre, face information, shot genre, image matching, and visual concepts. The detection of the first five concepts follows standard techniques as describe din Section 2.1. We will elaborate on the extraction of visual concepts using HMM here. To detect visual concepts within each keyframe image, we employ HMM to model the association between concepts and their positive images. This is done by segmenting each key-frame into fixed 4x4 blocks and followed by clustering them. The visual features used to represent each image block include color histogram, edge histogram, and the adaptive matching pursuit feature for texture. More specifically, they are: Luv color histogram, adaptive Matching Pursuit texture features and edge histogram (Shi et al, 2004). The dimension of the resulting visual feature vector is 90. Our method treats each block in a key-frame separately and a data point in low-level visual space is just a 90-dim block vector. We then perform k-means clustering (with k=500) to the block vectors of the training images. Each training image is tokenized by dividing the image into a set of regular-sized blocks, each represented by a feature vector, and quantized into a visual cluster. As shown in Figure 1 , the content representation of each image is modeled as having been stochastically generated by a HMM, where each state will generate some blocks with similar features. In addition, the spatial transactions and co-occurrences between fixed-size blocks within an image can be utilized to help in detecting the feature concepts. Figure 1: HMM for Concept i. Figure 2: Training and Testing Process for Visual feature. Q2 Q1

We introduce the challenge problem for generic video indexing to gain insight in intermediate steps that affect performance of multimedia analysis methods, while at the same time fostering repeatability of experiments. To arrive at a challenge problem, we provide a general scheme for the systematic examination of automated concept detection methods, by decomposing the generic video indexing problem into 2 unimodal analysis experiments, 2 multimodal analysis experiments, and 1 combined analysis experiment. For each experiment, we evaluate generic video indexing performance on 85 hours of international broadcast news data, from the TRECVID 2005/2006 benchmark, using a lexicon of 101 semantic concepts. By establishing a minimum performance on each experiment, the challenge problem allows for component-based optimization of the generic indexing issue, while simultaneously offering other researchers a reference for comparison during indexing methodology development. To stimulate further investigations in intermediate analysis steps that inuence video indexing performance, the challenge offers to the research community a manually annotated concept lexicon, pre-computed low-level multimedia features, trained classifier models, and five experiments together with baseline performance, which are all available at http://www.mediamill.nl/challenge/.

We have developed an interactive video search system that allows the searcher to rapidly assess query results and easily pivot off those results to form new queries. The system is intended to maximize the use of the discriminative power of the human searcher. This is accomplished by providing a hierarchical segmentation, streamlined interface, and redundant visual cues throughout. The typical search scenario includes a single searcher with the ability to search with text and content-based queries. In this paper, we evaluate new variations on our basic search system. In particular we test the system using only visual content-based search capabilities, and using paired searchers in a realtime collaboration. We present analysis and conclusions from these experiments.

We have developed an interactive video search system that allows the searcher to rapidly assess query results and easily pivot o those results to form new queries. The system is intended to maximize the use of the discriminative power of the human searcher. The typical video search scenario we consider has a single searcher with the ability to search with text and content-based queries. In this paper, we evaluate a new collaborative modication of our search system. Using our system, two or more users with a common information need search together, simultaneously. The collaborative system provides tools, user interfaces and, most importantly, algorithmically-mediated retrieval to focus, enhance and augment the team’s search and communication activities. In our evaluations, algorithmic mediation improved the collaborative performance of both retrieval (allowing a team of searchers to nd relevant information more eciently and eectively),

We combine in this paper automatic learning of a large lexicon of semantic concepts with traditional video retrieval methods into a novel approach to narrow the semantic gap. The core of the proposed solution is formed by the automatic detection of an unprecedented lexicon of 101 concepts. From there, we explore the combination of query-by-concept, query-by-example, query-by-keyword, and user interaction into the MediaMill semantic video search engine. We evaluate the search engine against the 2005 NIST TRECVID video retrieval benchmark, using an international broadcast news archive of 85 hours. Top ranking results show that the lexicon-driven search engine is highly effective for interactive video retrieval.

Video summarization refers to the process of recapitulating video stream by producing an abstract of the salient keyframes that could cover its overall content. However, an efficient video summarization requires an efficient video Shot Boundary Detection (SBD) and keyframes extraction. In this backdrop, this paper presents a novel and efficient approach for video SBD and keyframes extraction that will lead in the summarizing video. Meanly, Motion SIFT-Distribution Histogram (MoSIFT-DH) is extracted from the frames as a Glocal feature. The shot boundaries are detected using an adaptive threshold for the computed distance of MoSIFT-DH of the consecutive frames. Furthermore, keyframe representing the salient content of each segmented shot is extracted using entropy based singular values. Finally, the summarizing video is generated, by combining all the extracted keyframes. Our experiments on various videos indicate that our method can efficiently detect shot boundaries under different levels of illumination, camera operations and motion effects.

Video is the most challenging and has lots of issues than other type of multimedia data and video authentication is one of the popular issues regarding the video. Technologies of communication nowadays can simplify sharing of multimedia data such as text, images and videos through online or offline and a wide spread of video content its significant to assure the integrity and originality of video information. Recently, a lot of cases have been stated regarding an illegal activity in the video. Such as, video editing or video tampering attacks to ruin the authenticity of the video. With only minor alterations could change the meaning of the video. Video authentication detection is a popular issue and of high importance in various fields such as in forensic investigations, court of law and video surveillance. A lot of techniques are proposed by previous study to assure the originality of video in their own way. In this study, it focuses to detect tampering attacks and alterations in video using shot boundary detection techniques. Shot boundary detection is a very challenging task and it's commonly used for structured video because these techniques can detect the changes made with normal shot change rate. Meanwhile, this research will emerge to produce new model for video structure detection using shot boundary detection with the histograms feature that is significant to design of powerful tools to detect an originality of video and to locate the tampering attacks in video.

Video has become an interactive medium of daily use today. However, the sheer volume of the data makes it extremely difficult to browse and find required information. Organizing the video and locating required information effectively and efficiently presents a great challenge to the video retrieval community. This demands a tool which would break down the video into smaller and manageable units called shots. Traditional shot detection methods use pixel difference, histograms, or temporal slice analysis to detect hard-cuts and gradual transitions. However, systems need to be robust to sequences that contain dramatic illumination changes, shaky camera effects, and special effects such as fire, explosion, and synthetic screen split manipulations. Traditional systems produce false positives for these cases; i.e., they claim a shot break when there is none. We propose a shot detection system which reduces false positives even if all the above effects are cumulatively present in one sequence. Similarities between successive frames are computed by finding the correlation and is further analyzed using a wavelet transformation. A final filtering step is to use a trained Support Vector Machine (SVM). As a result, we achieve better accuracy (while retaining speed) in detecting shot-breaks when compared with other techniques.

Video concept detection aims to detect high-level semantic information present in video. State-of-the-art systems are based on visual features and use machine learning to build concept detectors from annotated examples. The choice of features and machine learning algorithms is of great influence on the accuracy of the concept detector. So far, intensitybased SIFT features based on interest regions have been applied with great success in image retrieval. Features based on interest regions, also known as local features, consist of an interest region detector and a region descriptor. In contrast to using intensity information only, we will extend both interest region detection and region description with color information in this thesis. We hypothesize that automated concept detection using interest region features benefits from the addition of color information. Our experiments, using the Mediamill Challenge benchmark, show that the combination of intensity features with color features improves significantly over intensity features alone.

Various query methods for video search exist. Because of the semantic gap each method has its limitations. We argue that for effective retrieval query methods need to be combined at retrieval time. However, switching query methods often involves a change in query and browsing interface, which puts a heavy burden on the user. In this paper, we propose a novel method for fast and effective search trough large video collections by embedding multiple query methods into a single browsing environment. To that end we introduced the notion of query threads, which contain a shot-based ranking of the video collection according to some feature-based similarity measure. On top of these threads we define several thread-based visualizations, ranging from fast targeted search to very broad exploratory search, with the Fork-Browser as the balance between fast search and video space exploration. We compare the effectiveness and efficiency of the ForkBrowser with the CrossBrowser on the TRECVID 2007 interactive search task. Results show that different query methods are needed for different types of search topics, and that the ForkBrowser requires significantly less user interactions to achieve the same result as the CrossBrowser. In addition, both browsers rank among the best interactive retrieval systems currently available.

University of Oulu’s MediaTeam research group participated to the manual and interactive search tasks with our video retrieval system, which is constructed out of three sub-applications: Content-based Query Tool, Cluster-Temporal Browser and Result Container with relevance feedback mechanism. The search engine in the heart of our search system creates result sets using three distinct search sub-engines that take care of the text, semantic concept and example based queries. The sub-applications and subengines result in varying search strategies. Therefore their effect on overall search performance has been studied in our experiments. The principal contribution in our search system is the Cluster-Temporal Browser, the effect of which we tested this year in conjunction with the traditional content-based paradigm: recurrent queries with relevance feedback. Seven search runs were submitted for the TRECVID 2005 search task: I_A_2_OUMT_I1Q_1: interactive with browsing disabled, expert users I_A_2_OUMT_I2B_2: interactive with browsing enabled, expert users I_A_2_OUMT_I3Q_3: interactive with browsing disabled, novice users I_A_2_OUMT_I4B_4: interactive with browsing enabled, novice users M_A_1_OUMT_M5T_5: manual with official text transcripts M_A_2_OUMT_M6TS_6: manual with official text transcripts and selected semantic concepts M_A_2_OUMT_M7TE_7: manual with official text transcripts and selected topic examples For interactive experiments, we have used eight test users: four novice and four expert users. In a carefully designed experimental setup, all users have used two system configurations: one with and one without Cluster-Temporal Browser. In our experiments we found that by incorporating Cluster-Temporal Browser the performance gain is more than 12% over the conventional content-based tools with novice users. We also found that on average expert users gained more than 18% better search performance over novice users, which shows that the test design has a significant effect to the outcome of the interactive test. Our manual experiment is a comparison of three query configurations: a baseline text only search, text combined with visual example search and text combined with semantic concept search. To facilitate a concept search, we trained a set of concept detectors (to name a few: newsroom, entertainment, maps and charts, news footage, weather, sports, outdoors). Our example search was based on low level descriptions (color and structure). The goal of the experiment was to find out how much more visual search examples and user defined semantic concepts contribute to the search performance over a text baseline. Our experimental results showed that the combined text and semantic concept search gives about 19% better performance over text baseline whereas text combined with example based search gives about 25% performance gain over the baseline. The results show that specific visual search examples accumulate better overall precision than the queries defined with our detected set of semantic concepts.

Traditional approaches for automatic video annotation usually represent one video clip with a flat feature vector, neglecting the fact that video data contain natural structures. It is also noteworthy that a video clip is often relevant to multiple concepts. Indeed, the video annotation task is inherently a Multi-Instance Multi-Label learning (MIML) problem. Considering that manually annotating videos is labor-intensive and time-consuming, this paper proposes a semi-supervised MIML approach, SSMIML, which is able to exploit abundant unannotated videos to help improve the annotation performance. This approach takes label correlations into account, and enforces similar instances to share similar multi-labels. Evaluation on TREVID 2005 show that the proposed approach outperforms several state-of-the-art methods.

Today's ubiquity of visual content as driven by the availability of broadband Internet, low-priced storage, and the omnipresence of camera equipped mobile devices conveys much of our thinking and feeling as individuals and as a society. As a result the growth of video repositories is increasing at enourmous rates with content now being embedded and shared through social media. To make use of this new form of social multimedia, concept detection, the automatic mapping of semantic concepts and video content has to be extended such that concept vocabularies are synchronized with current real-world events, systems can perform scalable concept learning with thousands of concepts, and high-level information such as sentiment can be extracted from visual content. To catch up with these demands the following three contributions are made in this thesis: (i) concept detection is linked to trending topics, (ii) visual learning from web videos is presented including the proper treatment of tags as concept labels, and (iii) the extension of concept detection with adjective noun pairs for sentiment analysis is proposed. In order for concept detection to satisfy users' current information needs, the notion of fixed concept vocabularies has to be reconsidered. This thesis presents a novel concept learning approach built upon dynamic vocabularies, which are automatically augmented with trending topics mined from social media. Once discovered, trending topics are evaluated by forecasting their future progression to predict high impact topics, which are then either mapped to an available static concept vocabulary or trained as individual concept detectors on demand. It is demonstrated in experiments on YouTube video clips that by a visual learning of trending topics, improvements of over 100% in concept detection accuracy can be achieved over static vocabularies (n=78,000). To remove manual efforts related to training data retrieval from YouTube and noise caused by tags being coarse, subjective and context-depedent, this thesis suggests an automatic concept-to-query mapping for the retrieval of relevant training video material, and active relevance filtering to generate reliable annotations from web video tags. Here, the relevance of web tags is modeled as a latent variable, which is combined with an active learning label refinement. In experiments on YouTube, active relevance filtering is found to outperform both automatic filtering and active learning approaches, leading to a reduction of required label inspections by 75% as compared to an expert annotated training dataset (n=100,000). Finally, it is demonstrated, that concept detection can serve as a key component to infer the sentiment reflected in visual content. To extend concept detection for sentiment analysis, adjective noun pairs (ANP) as novel entities for concept learning are proposed in this thesis. First a large-scale visual sentiment ontology consisting of 3,000 ANPs is automatically constructed by mining the web. From this ontology a mid-level representation of visual content – SentiBank – is trained to encode the visual presence of 1,200 ANPs. This novel approach of visual learning is validated in three independent experiments on sentiment prediction (n=2,000), emotion detection (n=807) and pornographic filtering (n=40,000). SentiBank is shown to outperform known low-level feature representations (sentiment prediction, pornography detection) or perform comparable to state-of-the art methods (emotion detection). Altogether, these contributions extend state-of-the-art concept detection approaches such that concept learning can be done autonomously from web videos on a large-scale, and can cope with novel semantic structures such as trending topics or adjective noun pairs, adding a new dimension to the understanding of video content.

This year, the University of Central Florida participated in the high level feature extraction task (HLF). The goal of high level feature extraction is to identify in videos specific shots that contain concepts such as “bus,” “person playing soccer,” and “boat/ship.” In our submissions, we focused on addressing the large imbalance between the positive and negative training examples. Specifically, we implemented a method called bootstrapping that identifies the best subset of negative examples to train on. In our experiments, we found bootstrapping significantly lowered the probability of false alarm while also improving the probability of detection. Additionally, we also explored different word weighting techniques. In the bag of words approach, certain words may be more discriminative than others; these words should be weighted more. This task served as a project for several students participating in the Research Experience for Undergraduates program (REU) at UCF.

This year, Indian Institute of Technology, Bombay participated in TRECVID 2006 in the task of shot detection. We observe that, even though large number of available shot detection methods perform well under normal conditions (as evinced by the results at Trecvid for the last couple of years), they are not robust to sequences that contain dramatic illumination changes, shaky camera effects, and special effects such as fire, explosion, and synthetic screen split manipulations. Traditional systems produce false positives for these cases; i.e., they claim a shot break when there is none. We propose a shot detection system which reduces false positives even if all the above effects are cumulatively present in one sequence. Similarities between successive frames are computed by finding the correlation and is further analyzed using a wavelet transformation. A final filtering step is to use a trained Support Vector Machine (SVM). As a result, we achieve better accuracy (while retaining speed) in detecting shotbreaks when compared with other techniques.