Domain Conversion for Cross-Domain Signal Quality Enhancement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing signal processing methods struggle to perform quality improvement processing effectively when the signal characteristics of an input signal differ from those of the student data used during learning, as they often lack suitable teacher data and cannot seamlessly switch between learning and inference.
Innovation Solution
A signal processing device that includes a quality improvement processing unit using a learning result from a first domain and a domain conversion unit to adapt input signals with different characteristics, utilizing deep neural networks for both units, along with a discriminator to identify signal domains and switch processing paths accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If quality improvement processing is performed using a learning result obtained from student data with specific signal characteristics, then processing accuracy is improved for signals with matching characteristics, but processing accuracy deteriorates for signals with different characteristics
Solution Approach 1:
The patent introduces a domain conversion unit as an intermediary component that converts input signals with different characteristics into the first domain before processing. This mediator enables the quality improvement processing unit to handle diverse signal types by transforming them into a standardized format that matches the training data characteristics, thereby resolving the contradiction between processing accuracy and signal adaptability
Solution Approach 2:
The patent changes the domain parameter of the input signal through the domain conversion unit. By transforming signals from different domains into the first domain, the system maintains consistent signal characteristics that match the training data, allowing the quality improvement processing to achieve high accuracy across various input signal types
2Adaptability or versatility
If learning is performed directly on unknown video signals or as online/background processing, then adaptability to different signal characteristics is improved, but system complexity increases due to the need for teacher data with matching content
Solution Approach 1:
The patent performs domain conversion as a preliminary action before quality improvement processing. By converting the input signal to the first domain in advance, the system prepares the signal in a format suitable for the trained model, eliminating the need for complex online learning mechanisms and teacher data matching while maintaining high adaptability
Solution Approach 2:
The patent segments the signal processing into two distinct functional units: a domain conversion unit and a quality improvement processing unit. This segmentation allows each unit to specialize in its specific task, with the domain conversion unit handling signal transformation and the quality improvement unit focusing on enhancement, thereby reducing overall system complexity
3Productivity
If a learning method is used without inference capability, then learning performance is improved, but continuous signal output is prevented requiring switching between learning and inference modes
Solution Approach 1:
The patent designs the quality improvement processing unit to serve dual purposes: it can perform inference on converted signals and also participate in learning processes. This multi-functionality allows the system to maintain continuous signal output capability while achieving high learning performance, as the same processing unit handles both operational modes without requiring mode switching
Data Source
AI summary
Quality improvement processing to which a learning result is applied can be performed appropriately even if the signal characteristics of an input signal are different from the signal characteristics of student data at the time of learning. A quality improvement processing unit applies quality improvement processing to an input signal to obtain an output signal using a learning result obtained using a signal of a first domain having first signal characteristics as student data. A domain conversion unit converts the input signal to a signal of the first domain and sends the same to a quality improvement processing unit when the input signal is a signal of a second domain having second signal characteristics different from the first signal characteristics.