Method for acquiring video scene relating value and video rapid browsing and searching method applying the method
A technology of video scene and acquisition method, applied in special data processing applications, character and pattern recognition, instruments, etc., can solve the problems of low accuracy and recall rate, neglect of scene correlation, etc., and achieve fast browsing and retrieval Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach 1
[0030] Specific implementation mode one: the following combination figure 1 Describe this embodiment, the acquisition method of video scene correlation value described in this embodiment, the step that realizes this method comprises:
[0031] Step a: preprocessing the video to be browsed, and extracting multiple key frames;
[0032] Step b: clustering the plurality of key frames using the K-means clustering algorithm, and initializing the correlation graph G between each video scene,
[0033] in, represents a collection of video scene categories, Represents the set of associated values between every two video scenes, initially empty;
[0034] Step c: let m represent the variable of the video scene, and initialize the variable m=1 of the video scene;
[0035] Step d: judging whether m satisfies m≤N, where N represents the number of video scenes;
[0036] If the judgment result is yes, execute the next step, if the judgment result is no, execute step 1,
...
specific Embodiment approach 2
[0062] Specific embodiment two: the video quick browsing method of the acquisition method of the video scene correlation value described in the application embodiment one described in this embodiment, the steps of realizing the method include: selecting the target video scene, and the N video The scenes are sorted from small to large according to the correlation value with the target scene, that is, the N video scenes are sorted from small to large according to the similarity with the target video scene, and then the quick browsing of the video scenes is completed.
[0063] The target video scene described in this embodiment is one of the interesting content in the video that the user wishes to quickly browse. Scene is an important component of video, which can be used as the basis for users to browse and search for videos. We first select a scene of interest, set it as the target video scene, and then, according to the degree of association between each video scene and the...
specific Embodiment approach 3
[0064] Specific embodiment three: the video retrieval method of the method for obtaining the associated value of the video scene described in the application embodiment one described in this embodiment, the steps for realizing the method include: selecting a target video scene, selecting a threshold T, and retaining the All video scenes with a correlation value greater than T between the N video scenes and the target video scene, that is, obtaining a video scene with a similarity greater than the threshold T with the target video scene, and the retrieval is completed;
[0065] After the retrieval function is completed, in order to facilitate browsing by the user, it may also include: sorting the retrieved video scenes according to the correlation value from small to large, that is, sorting according to the similarity with the target video scene.
[0066] In the actual application process, in addition to sorting according to the associated value, we can also choose to use ...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More