A video processing management platform based on image processing
A video processing and management platform technology, applied in image communication, selective content distribution, electrical components, etc., can solve problems such as large degrees of freedom, inability to achieve low latency, strong interactivity, and inability to customize and adjust, so as to reduce costs Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0032] This embodiment is one of the implementations of the video processing management platform of the present invention, as shown in the attached figure 1 , a video processing management platform based on image processing, the video processing management platform includes: a decoding module, which is used to input and decode a source video to be processed, so that the source video can generate a source that can be recognized and edited by a computer program Code; a modeling module, used to perform three-dimensional modeling on the target person in the source video, and generate an image model that can be edited by a computer program; an optimization module, used to receive the viewer's feedback on the source video and the target Character modification data and optimization opinions, thereby forming an optimization database and optimizing the target character, thereby obtaining an optimized code; an output module for integrating the optimized code into the source video, and re...
Embodiment 2
[0062] This embodiment should be understood as at least including all the features of any one of the foregoing embodiments, and further improvements on the basis thereof;
[0063] In the above embodiment, the various feature vectors E of the target person 1 ,E 2 ...E n Computational analysis is performed by machine learning, and a new feature vector E is generated through the mapping with the image model 1 ´, E 2 ´…E n ´To generate new images, and arrange the new images in time series to generate new video content; however, there are deficiencies in the algorithm based on machine learning, and due to the lack of a feature vector in the training data, the final generated feature may Vector E 1 ´, E 2 ´…E n ´The phenomenon of overfitting or unreasonable eigenvalues exists; at the same time, due to the insufficient description of the details of the image model, there are some unknown eigenvector changes after the adjustment of the target person, resulting in a lower-than...
Embodiment 3
[0071] This embodiment should be understood as at least including all the features of any one of the foregoing embodiments, and further improvements on the basis of it:
[0072] This embodiment includes presenting multiple fine-tuning schemes to multiple users, and after collecting evaluations from multiple users, adjusting the eigenvector optimization algorithm of the target person;
[0073] Related ways to expose to multiple users include:
[0074] 1. Select users to display, including selecting users at random; or,
[0075] 2. According to the degree of relevance between the feature vector involved in the fine-tuning scheme and the target display user, send it to the user with strong correlation, and use the Apriori association rule algorithm to compare multiple users and multiple target characters. Eigenvector E 1 ,E 2 ...E n Perform relevance calculations; including:
[0076] S1: extract the feature vector E whose support degree is greater than 20% as the frequent fe...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


