Image Processing Apparatus Noise Suppression Parabolic Approximation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing in mobile terminals struggles to extract detailed fingerprint or palm print information due to noise from JPEG compression and low-bit precision image sensors, which hinders differential operation for fine unevenness extraction.
Innovation Solution
An image processing apparatus that selects a rectangular sub-region around a target pixel, approximates the distribution of pixel values using a parabolic surface function via the least squares method, and calculates a normal vector from the approximation parameters to suppress noise and enhance unevenness extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If JPEG compression is applied to reduce image data size, then storage efficiency is improved, but noise increases and measurement precision deteriorates
Solution Approach 1:
The patent performs differential operation on the compressed image data before decompression to extract unevenness information. By conducting the differential calculation on the compressed JPEG data directly, the method avoids the noise amplification that would occur if decompression were performed first, thus preserving measurement precision while maintaining storage efficiency
Solution Approach 2:
The patent changes the processing approach by performing differential operation in the compressed domain rather than the decompressed domain. This parameter change in the processing sequence allows extraction of useful unevenness information while avoiding the noise amplification problem associated with decompressing high-compression JPEG images first
2Device complexity
If 8-bit precision image sensors are used to reduce device complexity, then manufacturing cost is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent extracts the useful unevenness information from the compressed image data by performing differential operation. This extraction process separates the meaningful surface unevenness information from the quantization noise, allowing high-precision fingerprint extraction even from 8-bit compressed images
Solution Approach 2:
The differential operation acts as an intermediary process that transforms the compressed image data into unevenness information. This intermediary transformation allows the system to overcome the limitations of 8-bit sensors by mathematically extracting the underlying surface characteristics that are otherwise obscured by compression noise
3Illumination intensity
If ISO sensitivity is increased to improve image capturing in low light, then illumination intensity is improved, but noise increases and measurement precision deteriorates
Solution Approach 1:
The patent performs differential operation on the compressed image data before any decompression or further processing. By conducting the unevenness extraction at this early stage, the method avoids amplifying the sensor noise that would be present in high-ISO images, thus maintaining measurement precision while allowing flexible ISO settings for different lighting conditions
Data Source
AI summary
An image processing apparatus comprises means for selecting, from a single image, a range where a target pixel is made to be a reference, means for approximating a distribution of pixel values of the selected range by a function that represents a curved surface, and means for calculating a vector related to the distribution of pixel values from a parameter of the function obtained as a result of the approximation.