Multiresolution Noise Estimation for Digital Imaging Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional imaging devices struggle to precisely and efficiently estimate noise levels in digital images, especially under varying imaging conditions, leading to inaccurate noise removal and image enhancement.
Innovation Solution
A method using a multiresolution model to estimate noise by capturing images under different conditions, applying multiresolution transformation, and building an a priori model database to generate noise level functions for each frequency layer, allowing for optimized noise estimation and subsequent image processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional noise estimation methods are used, then the process is simple, but the precision and efficiency of noise estimation is insufficient
Solution Approach 1:
The patent applies multiresolution transformation to segment the digital image into multiple frequency layers (low frequency, high frequency, very high frequency). This segmentation allows the noise estimation process to analyze different frequency components separately, improving precision by capturing noise characteristics at multiple resolution levels rather than treating the entire image uniformly.
Solution Approach 2:
The patent introduces a frequency domain dimension to the noise estimation process by transforming images from the spatial domain to the frequency domain using multiresolution transformation. This dimensional change enables analysis of noise characteristics in the frequency spectrum, providing additional information beyond what is available in the spatial domain alone.
2Productivity
If noise estimation is performed on entire digital images, then the process is straightforward, but the efficiency and precision is reduced due to varying noise levels across different frequency layers
Solution Approach 1:
The patent segments the image into different frequency layers through multiresolution transformation, allowing independent noise estimation for each layer. This segmentation enables the system to process only relevant frequency components, improving efficiency by avoiding unnecessary computation on uniform regions while maintaining high precision through layer-specific analysis.
Solution Approach 2:
The patent applies local quality analysis by estimating noise levels separately for different frequency layers and regions. Each frequency layer receives tailored noise estimation based on its specific characteristics, rather than applying a uniform noise model to the entire image. This local approach improves both precision (by accounting for spatially varying noise) and efficiency (by processing only necessary regions).
3Adaptability or versatility
If a priori model database is built with noise level functions from multiple imaging conditions, then the adaptability is improved, but the data processing complexity increases
Solution Approach 1:
The patent performs preliminary action by pre-building an a priori model database containing noise level functions from multiple imaging conditions before actual noise estimation is needed. This database is constructed in advance using training data, allowing the system to quickly query and apply appropriate noise models during actual image processing without performing complex analysis in real-time.
Solution Approach 2:
The patent creates a copy of noise characteristics from training images and stores them in the a priori model database. These copied noise level functions serve as reference models that can be applied to new images under similar conditions, reducing the need for real-time noise characterization and simplifying the actual estimation process while maintaining high adaptability.
Data Source
AI summary
A method for estimating noise according to a multiresolution model is applied to an imaging device and comprises steps of: using an imaging sensor of the imaging device to capture a series of images of a scene under different imaging conditions; processing the images with a multiresolution transformation process to obtain a series of sub-images corresponding to different frequency layers; processing a series of the sub-images of the images that are in a same frequency layer to generate an averaged image; determining a difference between each of the sub-images in the same frequency layer and the averaged image corresponding to that frequency layer, and calculating the differences and the averaged image to obtain noise level functions of the imaging sensor in the different frequency layers under the different imaging conditions; and defining the noise level functions of the imaging sensor as noise samples for establishing an a priori model database.