Compressed Image Sensing With Noise-Matched Reconstruction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing compressed sensing methods in image data reconstruction result in inhomogeneous overall images due to noise differences between sensed and reconstructed regions, leading to artifacts like stripes, and require excessive time and exposure, especially when imaging light-sensitive objects.

Innovation Solution

Reconstruct non-sensed image regions by transferring a noise signal from sensed regions to the reconstructed data, ensuring a homogeneous visual impression by using simulation models and noise estimation techniques, while preserving the original sensed data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If reconstruction methods predict reconstructed image data without noise signal, then image quality is improved, but inhomogeneous visual impression and stripe artifacts occur

Engineering Contradiction:

Improveimage qualityVSAvoidvisual homogeneity

Core Design Contradiction:

Measurement precisionVSStability of the object's composition

Solution Approach 1:

The patent applies different quality characteristics to different regions: sensed image regions maintain their original noise signal while reconstructed image regions are enhanced with artificial noise. This local differentiation resolves the contradiction by allowing high-quality reconstruction in non-sensed regions while maintaining visual homogeneity across the entire image through consistent noise characteristics.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent converts the harmful effect of noise absence in reconstructed regions (causing visual inhomogeneity) into a benefit by deliberately adding artificial noise to match the noise characteristics of sensed regions. This transforms the disadvantage of noise-free reconstruction into a solution that achieves both high image quality and visual homogeneity.

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

2Stability of the object's composition

If denoising methods are used to reduce noise signal, then homogeneous visual impression is achieved, but sensed image data are modified and no longer available as unfalsified data

Engineering Contradiction:

Improvevisual homogeneityVSAvoiddata integrity

Core Design Contradiction:

Stability of the object's compositionVSLoss of information

Solution Approach 1:

The patent segments the image into sensed regions and reconstructed regions, applying different processing strategies to each. Sensed regions are preserved unchanged to maintain data integrity, while reconstructed regions receive artificial noise to achieve visual homogeneity. This segmentation resolves the contradiction by protecting original data while still achieving the desired visual effect.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces artificial noise as an intermediary element that bridges the gap between noise-free reconstructed data and noisy sensed data. This intermediary noise signal allows visual homogeneity without modifying the original sensed data, thus preserving data integrity while achieving visual consistency.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Measurement precision

If very low to negligible noise signal is used, then image quality is improved, but great expenditure of time and high exposure are required

Engineering Contradiction:

Improveimage qualityVSAvoidexposure time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary action by generating artificial noise signals in advance and applying them to reconstructed image regions. This allows the system to achieve high image quality without requiring extended exposure times to capture sufficient signal, as the noise characteristics are pre-computed and applied computationally rather than acquired through prolonged exposure.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces the mechanical/physical approach of extending exposure time to reduce noise with a computational approach. Instead of physically exposing the sensor longer to capture cleaner data, the system uses algorithms to generate and apply artificial noise signals, substituting computational processing for physical exposure extension.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS12430720B2Method for compressed sensing and processing of image data

Publication Date: 2025.09.30 CARL ZEISS MICROSCOPY GMBH

AI summary

A method can be used for sensing and processing image data for an object to be imaged. The object is scanned incompletely by virtue of regions (eB) of the object being sensed, where the sensed image regions (eB) alternate with non-sensed image regions (neB) of the object. Image data (rBD) of the non-sensed image regions (neB) are reconstructed on the basis of the sensed image data (eBD) of the sensed image regions (eB). A noise signal (N) of the sensed image data (eBD) of the sensed regions (eB) is ascertained and transferred to the reconstructed image data (rBD) of the non-sensed regions (neB), so that a user obtains a homogeneous visual impression in relation to the noise arising in the overall image data of the object visualized in a resultant overall image (rGBInv).