Method and device for automatically dividing and classifying sports vision frequency shot
An automatic segmentation and lens technology, which is applied to color TV parts, TV system components, TVs, etc., can solve the problems of lack of versatility, inaccurate lens segmentation results, and lack of consideration of the particularity of sports video lenses. Achieve the effect of easy feature calculation and extraction
Inactive Publication Date: 2009-12-09
DONGGUAN LIANZHOU ELECTRONICS TECH
View PDF2 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
The shot segmentation of sports video can adopt the segmentation method of general video, and the boundary of the shot can be obtained through the similarity of adjacent frames, but the existing method does not consider the particularity of the shot of sports video, that is, the characteristics of fast movement and repetitive structure, so Shot segmentation results are not accurate
For the classification of sports video shots, some existing methods mainly use domain knowledge and specific rules, such as segmentation and classification of football video shots according to grass color ratio and personnel size. These methods can achieve good results for specific sports videos, but Not universal, unused sports videos have to derive different classification rules according to their characteristics
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0017] Sports videos can be divided into the following three types of shots, namely long-range shots, medium shots and close-up shots, such as figure 1 shown. The object of the present invention is to automatically segment and classify these three types of shots in sports videos. The present invention will be further described below in conjunction with the accompanying drawings.
[0018] The scheme of the present invention mainly includes the following steps: first divide the video into SSU sequences, extract color and motion features from each SSU, and then use the hidden Markov model training data to generate the hidden Markov model of each type of shot on the basis of feature extraction. The Koff model connects all possible shot model sequences into a shot network, and calculates the path with the largest logarithmic probability (log probability) from the shot network to obtain the final shot segmentation and classification results. The following is a detailed description...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention proposes a method for automatic segmentation and classification of sports video shots. This method divides the sports video stream into continuous Shot Sampling Units (SSU), and different shots correspond to different SSU timing sequences. Using the hidden Markov model to model different SSU sequences, the hidden Markov of different shots can be obtained. husband model. On this basis, all possible lens model combinations are concatenated to form a lens network. For each path in the lens network, that is, a sequence of lens models, calculate its log probability, and take the one with the highest probability as the best path, then all the lens models on the optimal path are the final classification results, and the corresponding SSU sequence The starting and ending SSUs are the boundaries of the shots, thus realizing the segmentation of sports video shots.
Description
technical field [0001] The invention relates to a video shot segmentation method and a device thereof, in particular to a method and a device for automatically segmenting and classifying sports video shots. Background technique [0002] In sports video, a shot is the basic structural unit of a sports video, and a sports video shot usually refers to a group of continuous image frames formed by shooting from a single camera perspective. Different types of sports video shots show different semantic content. For example, long-range shots usually reflect the overall situation of the game, while medium-range shots usually track the players’ movements, and close-up shots usually show the close-up of players and referees. Close-ups generally appear during the pause phase of the game. The shot segmentation of sports video can adopt the segmentation method of general video, and the boundary of the shot can be obtained through the similarity of adjacent frames, but the existing method...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/20H04N5/14
Inventor 杨颖林守勋张勇东
Owner DONGGUAN LIANZHOU ELECTRONICS TECH