Neural Network DeMux Mux Architecture for Image Enhancement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The efficiency of image enhancement using convolutional methods is relatively low, resulting in unsatisfactory enhanced images.
Innovation Solution
A neural network architecture comprising 2n sampling units and processing units, including DeMux and Mux units, and convolutional blocks, which rearranges and combines pixels following specific scrambling rules, and incorporates noise input to improve image enhancement. The network is trained using loss functions such as L1 loss, content loss, and adversarial loss to optimize image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional convolutional methods are used for image enhancement, then the enhancement process can be performed, but the computing efficiency is low and convergence speed is slow
Solution Approach 1:
The neural network is divided into multiple processing stages with DeMux units that split input images into multiple channels, process them in parallel, and then combine results through Mux units. This segmentation enables concurrent processing of different image components, significantly improving computing efficiency and reducing convergence time compared to traditional sequential convolutional methods.
2Manufacturing precision
If traditional convolutional methods are used for image enhancement, then the process can be completed, but the enhanced image quality is unsatisfactory
Solution Approach 1:
The patent transforms the traditional 2D convolutional processing into multi-channel parallel processing by dividing images into multiple processing channels through DeMux units. This dimensional transformation allows the network to process different frequency components and feature representations simultaneously, achieving superior image quality enhancement while maintaining high processing efficiency.
Data Source
AI summary
A neural network is provided. The neural network includes 2n number of sampling units sequentially connected; and a plurality of processing units. A respective one of the plurality of processing units is between two adjacent sampling units of the 2n number of sampling units. A first sampling unit to an n-th sample unit of the 2n number of sampling units are DeMux units. A respective one of the DeMux units is configured to rearrange pixels in a respective input image to the respective one of the DeMux units following a first scrambling rule to obtain a respective rearranged image. An (n+1)-th sample unit to a (2n)-th sample unit of the 2n number of sampling units are Mux units. A respective one of the Mux units is configured to combing respective m′ number of input images to the respective one of the Mux units to obtain a respective combined image.