Image processing method, model processing method, device and chip
By performing lightweight processing on the specified functional modules of the image segmentation model, the problem that the image segmentation model is difficult to deploy on terminal devices is solved, and image segmentation processing is realized when the network environment is poor, reducing the dependence on the network.
CN119722690APending Publication Date: 2025-03-28BEIJING XIAOMI MOBILE SOFTWARE CO LTD +1
0 Cites 0 Cited by
Patent Information
- Application Number
- CN202311279699.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-28
- Publication Date
- 2025-03-28
AI Technical Summary
Technical Problem
In the prior art, the image segmentation model is difficult to deploy on terminal devices due to the high demand for computing resources, resulting in the image segmentation process dependent on the network environment, especially when the network environment is poor.
Method used
By lightening the specified functional modules in the pre-trained image segmentation model, the target segmentation model can be obtained, so that it can be deployed on the terminal device to realize local processing of image segmentation.
Benefits of technology
It realizes that image segmentation processing can still be performed in the case of poor network environment, reduces dependence on the network environment, and enriches the functions of terminal devices.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention relates to an image processing method and device, a model processing method and device and a chip, and the method comprises the steps: inputting a to-be-segmented first image into a target segmentation model for segmentation after the first image is obtained, and obtaining a segmented second image. Compared with the prior art, the image segmentation process can be processed by the terminal itself, and it is not needed that in the prior art, after the first image is obtained, only the image is transmitted to the server side for segmentation processing, and the image segmentation result of the server side is received. Namely, according to the mode of the invention, after the terminal obtains the first image, the terminal can utilize the target segmentation model deployed by the terminal to perform segmentation processing on the first image, so that the segmentation processing of the image can be realized even under the condition that the network environment is relatively poor; therefore, the image processing process does not have strong dependence on the network environment as in the related technology, and functions which can be executed by the terminal are enriched at the same time.
Need to check novelty before this filing date? Find Prior Art