Image Compression Gain Control for Variable Bit Rates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image encoding methods based on deep convolutional networks can only output one encoding result for a specific type of input image, failing to achieve the desired compression bit rate according to actual requirements.
Innovation Solution
An image processing method that involves feature extraction, setting target compression bit rates, and performing quantization and entropy encoding to control the compression bit rate, using gain values and reverse gain values for encoding and decoding processes to achieve desired bit rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a deep convolutional network is used for image encoding, then encoding performance is improved, but the model can only output one encoding result for one type of input image, failing to achieve desired compression bit rate according to actual requirements
Solution Approach 1:
The patent introduces dynamic gain adjustment mechanisms where the encoding model can adaptively modify feature map values based on target compression bit rates. The gain values are dynamically determined according to the required compression level, allowing a single model to produce different encoding results for the same input image at different bit rates, thus resolving the contradiction between fixed encoding performance and variable bit rate requirements
Solution Approach 2:
The patent changes the parameter of gain values associated with feature maps. By adjusting these gain values as parameters, the model can transform the same feature extraction into different encoding outputs. The gain values are modified based on target compression bit rates, enabling the model to adapt to different compression requirements while maintaining the same underlying encoding architecture
Data Source
AI summary
An image processing method and apparatus are provided. The method includes: obtaining an image; performing feature extraction on the image to obtain at least one first feature map, wherein the at least one first feature map includes N first feature values, and N is a positive integer; obtaining a target compression bit rate which corresponds to M target gain values, each target gain value corresponds to one first feature value, and M is a positive integer less than or equal to N; respectively processing corresponding first feature values based on the M target gain values to obtain M second feature values; and performing quantization and entropy encoding on at least one processed first feature map to obtain encoded data, wherein the at least one processed first feature map includes the M second feature values. According to the application, compression bit rate control is implemented in a same compression model.