图像处理的方法、装置、介质
By jointly optimizing a video/image coding framework based on artificial neural networks, the problem of overall optimization of video codecs in existing technologies is solved, and the overall performance of image/video compression and rate distortion performance are improved.
CN117222997BActive Publication Date: 2026-07-17TENCENT AMERICA LLC
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TENCENT AMERICA LLC
- Filing Date
- 2023-03-23
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing video codecs are difficult to optimize as a whole, and improvements to individual modules cannot bring about overall performance improvements. The image/video compression process requires a lot of expertise and time.
Method used
We employ a video/image coding framework based on artificial neural networks, jointly optimize each module through a machine learning process to achieve end-to-end neural image compression (NIC), and utilize a multi-rate compression domain computer vision task decoder to perform computer vision tasks.
Benefits of technology
It achieves overall optimization of the image/video compression process, improves rate-distortion performance, and simplifies and accelerates the improvement of compression tools.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN117222997B_ABST
Abstract
在一些示例中，一种用于图像 / 视频处理的装置，包括处理电路。该处理电路根据携带有压缩图像的编码比特流，确定用于调整压缩图像的压缩率的参数的值。基于神经网络的编码器根据该参数的值生成压缩图像。该处理电路将该参数的值输入到多速率压缩域计算机视觉任务解码器中，多速率压缩域计算机视觉任务解码器包括一个或多个神经网络，一个或多个神经网络用于根据该参数的相应值从多个压缩图像执行计算机视觉任务，该参数的相应值用于生成压缩图像。多速率压缩域计算机视觉任务解码器根据编码比特流中的压缩图像和参数的值，生成计算机视觉任务结果。
Need to check novelty before this filing date? Find Prior Art