Shot Level Research Articles

Efficient Genre-Specific Semantic Video Indexing

Large video collections such as YouTube contain many different video genres, while in many applications the user might be interested in one or two specific video genres only. Thus, when users are querying the system with a specific semantic concept like AnchorPerson, and MovieStars, they are likely aiming a genre specific instantiation of this concept. Existing methods treat this problem as a classical learning problem leading to unnecessarily complex models. We propose a framework to detect visual-based genre-specific concepts in a more efficient and accurate way. We do so by using a two-step framework distinguishing two different levels. Genre-specific concept models are trained based on a training set with data labeled at video level for genres and at shot level for semantic concepts. In the classification stage, video genre classification is applied first to reduce the entire data set to a relatively small subset. Then, the genre-specific concept models are applied to this subset only. Experiments have been conducted on a small 28-h data set for genre-specific concept detection and a 4168-h (80 \thinspace031 videos) benchmark data set for genre-specific topic search. Experimental results show that our proposed two-step method is more efficient and effective than existing methods which do not consider the different semantic levels between video genres and semantic concepts for both the indexing and the search tasks. When filtering out 80% of the data set, the average performance loss is about 11.3% for genre-specific concept detection and 31.5% for genre-specific topic search, while the processing speed increases hundreds of times for different video genres.

VIOLENT SCENE DETECTION IN MOVIES

As violence in movies has harmful influence on children, in this paper, we propose a vision-based algorithm to detect violent scenes in movies. Under our definition of violence, the task of violent scene detection is decomposed into action scene detection and bloody frame detection. While previous approaches addressed only the shot level of video structure, our approach works on a more semantic-complete scene structure of video. The input video (digital movie) is first segmented into several scenes. Based on the filmmaking characteristics of action scene, some features of the scene are extracted to be fed into the support vector machine for classification. Finally, the face, blood and motion information are integrated to determine whether the action scene has violent content. Experimental results show that the proposed approach works reasonably well in detecting most of the violent scenes. Compared with related work, our approach is computationally simple yet effective.

Within and Between Shot Information Utilisation in Video Key Frame Extraction

In consequence of the popularity of family video recorders and the surge of Web 2.0, increasing amounts of videos have made the management and integration of the information in videos an urgent and important issue in video retrieval. Key frames, as a high-quality summary of videos, play an important role in the areas of video browsing, searching, categorisation, and indexing. An effective set of key frames should include major objects and events of the video sequence, and should contain minimum content redundancies. In this paper, an innovative key frame extraction method is proposed to select representative key frames for a video. By analysing the differences between frames and utilising the clustering technique, a set of key frame candidates (KFCs) is first selected at the shot level, and then the information within a video shot and between video shots is used to filter the candidate set to generate the final set of key frames. Experimental results on the TRECVID 2007 video dataset have demonstrated the effectiveness of our proposed key frame extraction method in terms of the percentage of the extracted key frames and the retrieval precision.

TOWARD ANALYZING THE IMPACT OF ADVERTISEMENT BILLBOARDS ON SOCCER TELECAST VIEWERS

Important soccer tournaments like the World Cup and the European Cup are broadcast to billions of people across the world. Therefore, advertising through billboards surrounding the playing arena in a soccer match is very attractive for promoting the brand image of a company. Analysis of strength and weakness of such an advertisement is helpful for the sponsors since a lot of money is spent on it and due to its potential impact on a large number of viewers at the same time. In this paper, we present a two-stage fuzzy system for analyzing the visual impact of advertisement billboards in soccer telecasts. In the first stage, visual impact of each shot is evaluated. Two parameters, namely, size and duration, of the billboards are considered in the fuzzy rules. After the shot level analysis, results are combined in the second stage to derive an overall visual impact. In both the stages, parameters of the fuzzy set membership functions are tuned using the Particle Swarm Optimization algorithm. The results have been compared against a user survey.

ICESat Full-Waveform Altimetry Compared to Airborne Laser Scanning Altimetry Over The Netherlands

Since 2003, the full-waveform laser altimetry system onboard NASA's Ice, Cloud and land Elevation Satellite (ICESat) has acquired a worldwide elevation database. ICESat data are widely applied for change detection of ice sheet mass balance, forest structure estimation, and digital terrain model generation of remote areas. ICESat's measurements will be continued by a follow-up mission. To fully assess the application possibilities of the full-waveform products of these missions, this research analyzes the vertical accuracy of ICESat products over complex terrain with respect to land cover type. For remote areas, validation of individual laser shots is often beyond reach. For a country with extensive geo-infrastructure such as The Netherlands, excellent countrywide validation is possible. Therefore, the ICESat full-waveform product GLA01 and the land elevation product GLA14 are compared to data from the Dutch airborne laser altimetry archive Actual Height model of the Netherlands (AHN). For a total population of 3172 waveforms, differences between ICESat- and AHN-derived terrain heights are determined. The average differences are below 25 cm over bare land and urban areas. Over forests, differences are even smaller but with slightly larger standard deviations of about 60 cm. Moreover, a waveform-based feature height comparison resulted in feature height differences of 1.89 m over forest, 1.48 m over urban areas, and 29 cm over low vegetation. These results, in combination with the presented processing chain and individual waveform examples, show that state-of-the-art ICESat waveform processing is able to analyze waveforms at the individual shot level, particularly outside urban areas.

Adaptive Edge-Oriented Shot Boundary Detection

We study the problem of video shot boundary detection using an adaptive edge-oriented framework. Our approach is distinct in its use of multiple multilevel features in the required processing. Adaptation is provided by a careful analysis of these multilevel features, based on shot variability. We consider three levels of adaptation: at the feature extraction stage using locally-adaptive edge maps, at the video sequence level, and at the individual shot level. We show how to provide adaptive parameters for the multilevel edge-based approach, and how to determine adaptive thresholds for the shot boundaries based on the characteristics of the particular shot being indexed. The result is a fast adaptive scheme that provides a slightly better performance in terms of robustness, and a five fold efficiency improvement in shot characterization and classification. The reported work has applications beyond direct video indexing, and could be used in real-time applications, such as in dynamic monitoring and modeling of video data traffic in multimedia communications, and in real-time video surveillance. Experimental results are included.

A video summarization tool using two-level redundancy detection for personal video recorders

Storage of video information is experiencing rapid growth as digital broadcasting services have developed. This has rapidly increased video data storage requirements in DVD and hard disc recorders in recent years. We design a video summarization tool for personal video recorders. This tool provides a video summarization function to grasp the original long video content quickly. This tool is with a two- level redundancy detection procedure. After the video parsing step, it removes redundant video content with the hierarchical agglomerative cluster method in the shot level. Impact factors of scenes and key frames are defined, and parts of shots are selected to generate the initial video summary. Finally a repetitive frame segment detection step is used to remove redundant information. We have developed a prototype model and tested this tool on TV series, movies and cartoons data. The experiment results are given to illustrate the proposed tool.

Content-trajectory approach for searching video databases

In the past few years, modeling and querying video databases have been a subject of extensive research to develop tools for effective search of videos. In this paper, we present a hierarchal approach to model videos at three levels, object level (OL), frame level (FL), and shot level (SL). The model captures the visual features of individual objects at OL, visual-spatio-temporal (VST) relationships between objects at FL, and time-varying visual features and time-varying VST relationships at SL. We call the combination of the time-varying visual features and the time-varying VST relationships a Content trajectory which is used to represent and index a shot. A novel query interface that allows users to describe the time-varying contents of complex video shots such as those of skiers, soccer players, etc., by sketch and feature specification is presented. Our experimental results prove the effectiveness of modeling and querying shots using the content trajectory approach.

Violence and consolation: September 11th 2001 covered by the Dutch children's news

The present paper describes a content analysis of the coverage of the events related to September 11th 2001 by the Dutch Children's News. Since 1980, the Dutch Broadcasting Corporation broadcasts a daily children's news programme (Jeugdjournaal) that is specifically designed for children aged 9–12. The Children's News selects news topics on the basis of their news value, which also implies coverage of violence and crime. To offset possible harmful effects of a violent news topic, the Children's News usually applies several ‘consolation strategies’. These strategies may be applied at the programme level (e.g. alternating ‘heavy’ and ‘light’ topics), at the level of the news item (e.g. taking a child perspective, inclusion of reassuring comments), and at the shot level (e.g. not showing the most graphic film footage). In the present study, it was investigated to what extent the Children's News applied these consolation strategies in the first week after the attack on the World Trade Center in New York.

A Multi-Modal Approach to Story Segmentation for News Video

This research proposes a two-level, multi-modal framework to perform the segmentation and classification of news video into single-story semantic units. The video is analyzed at the shot and story unit (or scene) levels using a variety of features and techniques. At the shot level, we employ Decision Trees technique to classify the shots into one of 13 predefined categories or mid-level features. At the scene/story level, we perform the HMM (Hidden Markov Models) analysis to locate story boundaries. Our initial results indicate that we could achieve a high accuracy of over 95% for shot classification, and over 89% in F1 measure on scene/story boundary detection. Detailed analysis reveals that HMM is effective in identifying dominant features, which helps in locating story boundaries. Our eventual goal is to support the retrieval of news video at story unit level, together with associated texts retrieved from related news sites on the web.

GammaPlan-Leksell Gamma Knife radiosurgery treatment planning verification method.

This work provides a method for an independent check of Gamma Knife GammaPlan radiosurgery calculations, named the spherical approximation method or SAM. Based on skull dimension measurements, the treated volume of the head is modeled as a sphere of radius R. With this approximation, an analytical solution for fast ray tracing of the path length, for each of the 201 beamlets, of the Gamma Knife helmet collimator was possible. The dose rate at the focus of a single shot is the sum of the contributions of all active beamlets adjusted for both the collimator factor and attenuation. For an arbitrary point, the dose rate is derived at the beamlet level from the focus values adjusted for the new path length attenuation and the appropriate collimators' off-axis profiles. The sum over all beamlets' contributions gives the dose rate at that particular point. At the single shot level, SAM independent check results agree with the GammaPlan for patient calculations to better than +/-6% and, as expected, in spherical phantoms the agreements improve to better than +/-1.0%. For an arbitrary point, multi-shot procedure, the agreement is better than +/-3% and +/-1.5, respectively.

Single Shot Unit Spray Testing of Metered Dose Inhalers

An apparatus has been developed for the collection of single shots of drug delivered through the actuator of metered dose inhalers. The equipment is constructed from readily available components and affords complete drug recovery in 50 ml solvent. The apparatus has been evaluated with three drugs commonly administered from metered dose inhalers; beclomethasone dipropionate, salbutamol and sodium cromoglycate. Assays at the single shot level are highly reproducible, with coefficients of variation of typically 5 %. The use of the apparatus provides a direct measure of the dose of drug available for inhalation by the patient.

Kinematic analysis of strong motion P and SV waves from the Sterling Event

P and SV velocity waveforms from the Sterling explosion of December 3, 1966, are analyzed from a point of view often taken in earthquake source studies. The detonation occurred in the cavity excavated by the Salmon explosion within the Tatum salt dome, Mississippi. Velocity recordings taken from nearby borehole strong motion instruments show the occurrence of significant SV energy (Perret, 1968a, b). Kinematic source models are constructed by considering constraints afforded by the preshot cavity, condition of the surrounding material, and aspects of the P and SV waveforms such as wave polarity and duration. Waveforms for these models are computed and compared directly with the data. Modeling results suggest that the SV waves were radiated primarily by induced normal faulting occurring beneath the cavity rather than by an asymmetric pressure distribution based on the preshot cavity geometry. Secondary arrivals within the P waveforms and SV amplitudes suggest that these normal faults also had a component of tensional displacement which was probably due to forceable injection of material from the blast. A measurement for Qβ of 35 is obtained for salt from the amplitude decay arid pulse width changes observed for SV waves along a profile taken at shot level. This value for Qβ is for the region away from the cracked area induced by the previous Salmon explosion and qualitatively explains the differing frequency content observed at the closest Sterling sensors between P and SV waves. The anomalous SV radiation has little to do with ‘tectonic’ stress release but is a result of the medium's geometry and anisotropy.

Seismic source energies of four explosions in a salt dome

Free-field ground motion induced by four explosions within the Tatum salt dome in Mississippi was recorded by a single set of six shot level instrument stations installed within the dome. These explosions ranged from a 5.3-kT tamped nuclear event to a 2.7-ton nitromethane detonation. Energies associated with free-field motions have been calculated, and those derived from data considered to have been from within the seismic region were compared with the energy released by each explosion. Ratios of seismic source energies to explosive energies are Salmon, ≤3.3%; Sterling, 0.02%; Sterling HE, 1.2%; and Diode Tube, 0.0015%. Comparison of these ratios indicates (1) that the Sterling decoupled nuclear event was in fact decoupled by a factor of about 100–170 from the fully coupled Salmon nuclear explosion; (2) that the seismic source energy of the tamped chemical explosion, Sterling HE, was roughly equivalent to that of tamped nuclear explosions in hard rock; and (3) that the seismic source energy of Diode Tube, the detonable gas explosion designed to be comparable to the Sterling nuclear explosion, was in fact lower than that of Sterling by a factor of nearly 15.

GHOST REFLECTIONS CAUSED BY ENERGY INITIALLY REFLECTED ABOVE THE LEVEL OF THE SHOT

Some experiments are described which demonstrate the occurrence of an initial reflection from above shot level and of a resultant secondary reflection from depth which, by optical analogy, may be termed a ghost reflection. The reflector above shot level was in this case the base of the weathered layer. Reflection coefficients as high as [Formula: see text] were found in the Pleistocene Beaumont clay surface in Harris County, Texas. With reflection coefficients of this magnitude, the ghost reflections influence the character of seismograms noticeably.