Adaptive Noise Criterion for Image Sensor Texture Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image sensors face challenges in effectively reducing noise, as indiscriminate noise determination criteria can mistakenly identify texture information as noise, degrading image quality and requiring a method that can variably determine noise based on image characteristics.
Innovation Solution
A method and apparatus that calculate color characteristic values for image blocks, compare them to a preset initial noise criterion, sort and accumulate results to modify the noise criterion, allowing for differentiated noise determination based on image characteristics, thereby distinguishing noise from texture information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If indiscriminate noise determination criteria are applied, then noise reduction is achieved, but texture information is erroneously removed degrading image quality
Solution Approach 1:
The patent applies different noise determination criteria to different blocks based on their local color characteristics. By calculating color characteristic values for each block and comparing them with initial noise criteria, the system adapts the noise determination threshold locally rather than applying a uniform criterion globally, thus preserving texture information while removing noise.
Solution Approach 2:
The noise determination criteria are dynamically adjusted based on accumulated comparison results from multiple blocks. The system modifies the initial noise criteria using accumulation results to generate adapted criteria that reflect the actual image characteristics, making the noise reduction process adaptive rather than static.
2Manufacturing precision
If variable noise determination is implemented, then image quality is preserved, but device complexity increases
Solution Approach 1:
The patent divides the image into multiple blocks and processes each block independently by calculating color characteristic values for each block. This segmentation allows the complex variable noise determination to be broken down into simpler, manageable unit operations that can be performed on individual blocks rather than the entire image at once.
Solution Approach 2:
The system uses the image's own color characteristic values to determine noise criteria for each block. By comparing color characteristics with initial noise criteria and using accumulation results to adapt the criteria, the algorithm makes the noise determination self-contained for each block, reducing the need for external complex processing.
3Measurement precision
If color characteristic calculation and accumulation is performed, then noise differentiation accuracy is improved, but memory requirements increase
Solution Approach 1:
The patent accumulates comparison results separately for each color characteristic class (first group, second group, etc.) rather than maintaining all raw data. By segmenting the accumulation by color characteristic categories, the system achieves accurate noise differentiation while limiting memory usage to only the necessary aggregated statistics for each class.
Data Source
AI summary
A method and an apparatus are provided for estimating noise determination criteria in an image sensor. The method includes calculating a color characteristic value for each of a plurality blocks constituting an input image, comparing the color characteristic value of a first block among the blocks with a initial noise criterion, sorting the color characteristic value of the first block as a first group of a color characteristic class and accumulating a result of the comparing into the first group, and modifying the initial noise criterion using a result of the accumulating and calculating a first group noise criterion to be applied to corresponding blocks belonging to the sorted first group of the color characteristic class.