Method, device, system, electronic device, and storage medium for image processing

An image processing method, characterized in that, applied at a decoding end, the method comprises:

• receiving a first code stream and a second code stream from an encoding end;
• inverse-quantizing and decoding the first code stream to obtain a one-dimensional feature vector;
• inverse-discretely decoding the second code stream to obtain a multi-dimensional feature map;
• reconstructing the image based on the one-dimensional feature vector and the multi-dimensional feature map to obtain a reconstructed image block;
• wherein the step of reconstructing the image based on the one-dimensional feature vector and the multi-dimensional feature map includes iteratively denoising the multi-dimensional feature map multiple times using the one-dimensional feature vector through a preset diffusion model, thereby obtaining the reconstructed image block.

An image processing system, characterized in that the system comprises:

• an encoding end, configured to execute an encoding image processing method, the encoding image processing method comprises:
• extracting a one-dimensional feature vector from an original image block;
• transforming the original image block into a multi-dimensional feature map based on the one-dimensional feature vector;
• quantizing and encoding the one-dimensional feature vector to generate a first code stream;
• discretely encoding the multi-dimensional feature map to generate a second code stream;
• sending the first code stream and the second code stream to a decoding end;
• wherein the step of transforming the original image block into a multi-dimensional feature map based on the one-dimensional feature vector includes:
• iteratively adding noise to the original image block multiple times using the one-dimensional feature vector through a preset diffusion model, thereby obtaining the multi-dimensional feature map and;
• a decoding end, configured to execute a decoding image processing method, the decoding image processing method comprises:
• receiving a first code stream and a second code stream from the encoding end;
• inverse-quantizing and decoding the first code stream to obtain a one-dimensional feature vector:
• inverse-discretely decoding the second code stream to obtain a multi-dimensional feature map;
• reconstructing the image based on the one-dimensional feature vector and the multi-dimensional feature map to obtain a reconstructed image block;
• wherein the step of reconstructing the image based on the one-dimensional feature vector and the multi-dimensional feature map includes iteratively denoising the multi-dimensional feature map multiple times using the one-dimensional feature vector through a preset diffusion model, thereby obtaining the reconstructed image block.