R. Zabih, J. Miller, and K. Mai. A feature-based algorithm for detecting and classifying scene breaks. ACM Multimedia, 1995. 2  Home/Search   Document Details and Download   Summary   Related Articles   Check

....shot transition dissolve. 3.1.1 Cut Detection The di#erence in grey level as well as the colour information between two consecutive frames is usually large at an abrupt shot boundary due to the content dissimilarity of the two shots. Many of the early methods for cut detection [ZKS93, HJW94b, ZMM95, AL96] were based on di#erence metrics, such as pixel intensity value di#erence and histogram di#erence. One of the problems with these di#erence based algorithms is that they are sensitive to busy scenes, in which intensities change substantially from frame to frame due to camera object motion. ....

....moves substantially from right to left during the dissolve. Figure 5 in Paper 1 shows the response of the proposed method for the video sequence containing this dissolve (the first peak) The result clearly shows that the proposed method is able to reliably detect this dissolve. Zabih et al. ZMM95] compared the results of several pixel domain algorithms applied to this video sequence and concluded that only their feature based algorithm can detect this kind of dissolves. Nevertheless, their algorithm operates on pixel data only and requires several computationally intensive steps (Gaussian ....

R. Zabih, J. Miller, and K. Mai. A feature-based algorithm for detecting and classifying scene breaks. In Proc. of ACM Multimedia, pages 189--200, San Francisco, CA, 1995.

.... No [13] with histogram local maxima comparison Shahraray [14] Graylevel thresholds Yes differences motion analysis Xiong al. Resolution Intensity Given threshold No [15, 16] or color reduction differences Zabih al. Edge Edge Given threshold Yes [17, 18] detection comparison local maxima Zhang al. thresholds Yes [19] or color comparison Our Graylevel Partial thresholds Yes [10] derivatives Table 1: Comparative table of methods using non compressed video for detection of shot boundaries. a series of ....

....choice of sampling ratio, the impact on the accuracy of the algorithm is below a reasonable limit while the computational gain is considerable. 4 Performance Evaluation 4. 1 Evaluation Method The performance of our algorithm comparing to two other (region histograms [11] and edge tracking [17, 18], summarized in Section 2) will be presented using two separate video sequences. While our choice of methods for comparison is arbitrary, both these methods have well performed when tested in survey papers [3, 5] For each implemented algorithm, the number of shot boundaries correctly detected, ....

[Article contains additional citation context not shown here]

R. Zabih, J. Miller, and K. Mai. A Feature-based Algorithm for Detecting and Classifying Scene Breaks. Technical Report CS-TR-95-1530, Computer Science Department, Cornell University, 1995.

....component in available multimedia data. We must now address the indexing problem emerging from the need to manage these video documents. Classical text based indexing methods are insufficient to provide an adequate description, so a new form of indexing is needed for video sequences. Many authors [5, 7, 11, 17, 19] believe that shot boundary detection in video sequences is one of the necessary first steps in an efficient video management system. Segmentation of digital video into smaller units is also important in other domains like MPEG compression. We define the digital video segmentation problem as the ....

....propose to apply their method to each band separately. Some authors [12, 19] suggest to reduce color space to obtain a limited number of different colors. Finally Lee and Ip [11] use HSI space by conserving band H and S because they represent color independently from the intensity. Zabih et al. [17] take contour information to identify cut boundaries. This kind of approach is rarely used because it is time consuming. To reduce computational time cost and because there is some redundancy in spatial and temporal information, many authors [1, 2, 15] propose to sample spatially and or ....

[Article contains additional citation context not shown here]

R. Zabih, J. Miller, and K. Mai. A Feature-based Algorithm for Detecting and Classifying Scene Breaks. In Proceedings of the ACM International Conference on Multimedia, pages 189--200, San Francisco, 1995. 16

....found by computing an image based distance between consecutive frames of the video. Over a certain threshold we consider that there is a cut. The distance between frames can be based on statistical properties of pixels [8] histogram difference [9] compression algorithms [10] edge differences [11] or motion detection [12] We use an automatic shot boundary detection developed in our laboratory [13 14] And we automatically annotate the camera motion. Using the resulting video segmentation, we annotate each shot with keywords and a degree of certainty. We use an annotation program ....

. Zabih, R., Miller, J., Mai, K., "A Feature-based Algorithm for Detecting and Classifying Scenes Breaks", Proc. ACM Multimedia 95, 189-200, San Francisco, CA, November 1995.

....the screen, with the new scene appearing behind the line. 2. History and References in Regards to This Project This project was initiated by Aidan Totterdell, who designed a program, edge adaptive, to detect cuts using the method of edge detection [1] The algorithm is based on that discussed in [2]. The project was then continued as a final year project, where it was rigorously tested against an eight hour indexed baseline of day time Television, and a report written documenting a detailed description of the algorithm used, the code written, the test results and the characteristics of ....

....within the strip should be higher than those in the rest of the image. The location of the changing pixels can be recorded and their spatial distribution analysed. 13 An alternative method could be to use motion vectors to recognise the wipe. However, this will need to be researched in more detail. [2] 6. Testing As can be seen in section 5.3.4, many different methods were tried and tested to improve the accuracy with which dissolves are detected. Unfortunately, no improvement was found on the initial method used. This method detects approximately 70 of the dissolves in a twenty minute ....

[Article contains additional citation context not shown here]

R.Zabih, J. Miller, K. Mai "A feature based algorithm for detecting and classifying scene breaks", Cornell University, USA.

....constant image (fade out) or between a constant image and a scene (fade in) a dissolve is a gradual transition from one scene to another, in which the first scene fades out and the second fades in. a wipe occurs as a line moves across the screen, with the new scene appearing behind the line. [3] There are many different types of scenes that can occur in broadcast video. As seen already, to detect shot boundaries, cuts, fades, dissolves and wipes need to be detectable. Modern television productions make extensive use of effects such as: Computer generated effects: This can be ....

....a few frames, and a fade out which looks like half of a dissolve. The edge change fraction for a dissolve is not as high as a cut as, pixels enter and exit gradually. Figure 4.1 Plot of video sequence demonstrating edge change fraction values in relation to a cut, dissolve and wipe, respectively. [3] Figure 4.1 Plot of video sequence demonstrating edge change fraction values in relation to a cut, dissolve and wipe Wipes can be distinguished from dissolves by looking at the spatial distribution of entering and exiting pixels. During a wipe each frame will have a portion of the old scene and ....

R.Zabih, J. Miller, K. Mai "A feature based algorithm for detecting and classifying scene breaks", Cornell University, USA.

....were chosen for test purposes. The trade off in using a higher resolution size against a lower one is the accuracy of results in the edge change fraction versus the increased execution time of the computation. The method we use to compute the edge change fraction, p, is based on that used in [4] to compare every pixel in the first undilated frame, E, against the corresponding pixels in the second dilated frame, E . There are two possibilities for this comparison: if a pixel is found in location (x,y) in frame E, and a matching pixel is found in its dilated area (x dx,y dy) in the ....

R. Zabih et al. "A Feature Based Algorithm for detecting and Classifying Production Effects", Multimedia Systems, Vol 7, pp119-128, 1999.

....comparison of corresponding pixels in adjacent frames; if more than a certain proportion of pixels are judged to have changed between frames, a scene change is declared at that point. Kasturi and Jain [2] proposed a method based on the mean and variance of intensities of image regions. Zabih [3] used edge detection and examined the number of entering and exiting edges between frames to determine whether or not a scene change had taken place. Probably the most widely used algorithms are based on colour histograms, where the colour space of the image is quantised and each pixel in the ....

Zabih R, Miller J, Mai K. A Feature-Based Algorithm for Detecting and Classifying Scene Breaks. Proceedings of 3rd ACM International Conference on Multimedia, Nov 1995.

....the problem, pixel differences, statistical differences, histograms, compression differences, edge tracking or motion vector differences. A comparison between these approaches can be found in [1] The most common approach to obtain boundaries of a shot is based on color information of the frame, [4, 8, 2, 9]. In the recent studies, no matter which color space is of interest, main advantages of using color is its ease of implementation, descriptive characteristic both in spatial and temporal space and real time applicability due to simplicity to obtain a feature vector, such as histogram [7] However, ....

....data, they can t be applied in real time environments. Edge map based approaches for shot detection use the assumption that the changes in spatial domain will result in appearing or fading of edges. The percentage of change in edges decides the shot boundary. Though the approach, as stated in [8], is sensitive to motion, the response time of the algorithm is not real time, because of the motion compensation step, thus can t be applied to surveillance systems. To sum up, all algorithms in the printed literature suffer the trade of between speed and accuracy. On the other hand, it should ....

R. Zabih, J. Miller, and K. Mai, "A feature-based algorithm for detecting and classifying scene breaks," Proc. ACM Multimedia 95, San Francisco, CA, November, 1993, pp. 189-200.

....8 x 8 block, they use the ideal step edge model and derive approximate edge orientation, offset, and strength parameters using only DCT AC coefficients. The result is coarse edge segments that can be used for video indexing tasks like shot boundary detection based on the change ratio of edge maps [67]. They also investigate convolution based edge detection in compressed images by merging symmetric kernel convolution with the IDCT procedures [48] Compared with conventional convolution based edge detection, a speedup of 3 to 10 times is achieved with comparable results. However, large ....

....paper, Lienhart [35] compares four major shot boundary detection algorithms, which include fade and dissolve detection. Extensive experimental results also favor the color 29 histogram based method [8] for shot boundary detection, instead of the computationally expensive edge change ratio method [67]. Researchers have also investigated compressed domain shot boundary detection techniques. One obvious approach is to apply uncompressed domain techniques to DCT DC image sequences using approximated features. Techniques using pixel difference, intensity statistics, and histograms are still ....

R. Zabih, J. Miller, and K. Mai, "A feature-based algorithm for detecting and classifying scene breaks," Proc. ACM Multimedia 95, pp. 189-200, Nov. 1995, San Francisco, CA.

....noisereduction techniques. The selected region tracks a simplified background area of a video frame. As in BGT, the ra tionale for selecting this area is to prevent scene cuts due to object movements or camera movements such as tilting and panning. Other techniques such as Edge Change Ratio (ECR) [11] are more sensitive to these movements and generate more false cuts in general. For each frame, pixel values of the region are extracted and transformed into a two dimensional array of pixel values, called TBA, as shown in Figure 1(b) The width and height of the TBA are bw and bh in pixels, ....

R. Zabih, J. Miller, and K. Mai. A feature-based algorithm for detecting and classifying scene breaks. In Proc. of ACM Multimedia'95, pages 189--200, San Francisco, CA, November 1995.

....If a large percentage of the blocks had changes in the image gradient greater than some threshold, the presence of a dissolve or fade was confirmed. The same technique was used to locate the start and end of each gradual transition. This approach was based on the work of Zabih, Miller, and Mai [17], the major differences being that we did not perform motion compensation and used raw image gradients rather than edges. This resulted in a method that was less computationally expensive than the typical edge entering and exiting method. The method did not perform well in the evaluation, but we ....

R. Zabih, J. Miller, K. Mai. `Feature-based Algorithms for Detecting and Classifying Scene Breaks.' Proceedings of the 4th ACM Int. Conf. on Multimedia, 1995.

....at a similar false alarm rate. Correct False Alarms Basic 0.70 0.40 Alomerative 0.81 0.41 Table 1: Results of basic and agglomerative audio segmentation schemes on 30 minutes of a TV news recording. cuts, fades, dissolves and wipes by the pattern of feature appearance or disappearance [11]. This is around 5(100 times slower than histogram comparison. We have implemented a regional colour histogram method, which represents a good tradeoff between speed and accuracy. We introduce a novel multi timescale filter bank to detect and distinguish between cut, dissolve and fade effects ....

Zabih, R., Miller J., and Mai, K., A Feature-Based Algorithm for Detecting and Classifying Scene Breaks, Proc. ACM Multimedia 95, San Francisco, CA, 1995.

....analog video cameras, this information either does not exist or is not externally available. If the video is MPEG encoded by an external device, any time information even in DV is gone. Therefore, shot detection is performed on the visual stream. Shot detection is a widely researched area [1 6]. We use the hard cut detection algorithm proposed in [5] and the fade detector in [4] to determine shots. Since both detectors work on only the DC coefficients of the MPEG video, our shot detector runs 19 times faster than real time on a Pentium II 450 MHz. Once the shots have been detected, a ....

R. Zabih, J. Miller, and K. Mai. A Feature-Based Algorithm for Detecting and Classifying Scene Breaks. Proceedings ACM Multimedia 95, San Francisco, CA, pp. 189-200, Nov. 1995.

No context found.

R. Zabih, J. Miller, and K. Mai. A feature-based algorithm for detecting and classifying scene breaks. ACM Multimedia, 1995. 2

No context found.

R. Zabih, J. Miller, and K. Mai. A feature-based algorithm for detecting and classifying production effects. Multimedia Systems, 7(2):119--128, 1999.

No context found.

R. Zabih, J. Miller, and K. Mai, "A feature-based algorithm for detecting and classifying scene breaks," in Proceedings ACM International Conference on Multimedia, pp. 189--200, ACM Press, 1995.

No context found.

R. Zabih, J. Miller, and K. Mai, "A feature-based algorithm for detecting and classifying production effects," Multimedia Systems, vol. 7, no. 2, pp. 119--128, 1999.

No context found.

R. Zabih, J. Miller and K. Mai, "A feature-based algorithm for detecting and classification production effects", Multimedia Systems, vol. 7, pp. 119-128, 1999.

No context found.

R. Zabih, J. Miller and K. Mai, "A feature-based algorithm for detecting and classifying scene breaks", Proc. ACM Multimedia `95, San Francisco, CA, pp. 189-200, Nov. 1995.

No context found.

R. Zabih, J. Miller, and K. Mai, "A feature-based algorithm for detecting and classifying production effects," Multimedia systems, vol. 7, pp. 119--128, 1999.

No context found.

R. Zabih, J. Miller and K. Mai, Feature-based algorithms for detecting and classifying scene breaks. Proceedings of the Third ACM Conference on Multimedia, pp 189-200, San Francisco, CA, November 1995, (with).]

No context found.

R. Zabih, J. Miller, K. Mai, "A feature-based algorithm for detecting and classifying production effects," Multimedia Systems, vol. 7, (no.2), Springer-Verlag, p. 119-28, March 1999.

No context found.

Zabih, R., Miller, J., and Mai, K., "A Feature-Based Algorithm for Detecting and Classifying Scene Breaks", Proc. ACM Multimedia 95, San Francisco, CA, November, 1993, pp. 189-200.

No context found.

Ramin Zabih, Justin Miller, and Kevin Mai. A feature-based algorithm for detecting and classifying scene breaks. In 3rd International Multimedia Conference and Exhibition, Multimedia '95, pages 189#200. ACM, New York, NY, USA, 59November

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC