An image processing method based on mirror edge filling and cutting and related components
By using mirroring and cropping techniques, the problems of edge distortion and size mismatch in block inference of large-size images are solved, thereby improving the quality and consistency of image processing.
CN122134873APending Publication Date: 2026-06-02深圳牛学长科技有限公司
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 深圳牛学长科技有限公司
- Filing Date
- 2026-04-13
- Publication Date
- 2026-06-02
AI Technical Summary
Technical Problem
When large images are used for block-based inference, the lack of neighboring pixel information in edge blocks leads to blurring and distortion, and the output image does not match the size of the original image, increasing processing complexity.
Method used
The image edge patching method is used to patch the edges of the image by dividing it into blocks. It uses the pixel information of the image itself to supplement the edges, and performs a cropping operation after fusion to remove redundant pixel areas, ensuring that the output image matches the size of the original image.
Benefits of technology
It improves the processing quality of edge areas, enhances the overall consistency and size matching of output images, and reduces the need for subsequent manual cropping.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122134873A_ABST
Abstract
This invention provides an image processing method and related components based on mirrored edge patching and cropping. The key technical points are: acquiring the original image to be processed; calculating the number of blocks and the patching size based on the block size and the overlapping area size; performing outward mirrored edge patching on the four edge blocks of the image to obtain the patched image; performing block segmentation and model inference on the patched image to obtain multiple inferred sub-blocks; weighted fusion of all sub-blocks to obtain the fused image; and cropping the fused image based on the original image size and the model magnification factor to remove redundant pixels generated by the edge patching, resulting in a final output image matching the original image size. Compared with existing technologies, this invention eliminates edge effects through mirrored edge patching and restores the original size through cropping, thus improving the matching degree between the output image and the original image size.
Need to check novelty before this filing date? Find Prior Art