Microscopic Image Focus Assessment With Gradient-Image Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the focusing quality of microscopic images of biological cellular substrates are prone to errors due to the presence of objects like particles or bubbles, leading to incorrect focusing planes and subsequent flawed pattern detection.
Innovation Solution
A method using a neural network to analyze both gradient and microscopic images simultaneously, determining a focusing measure by processing partial images from both to ensure accurate alignment with the cellular substrate plane.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional focusing methods are used to determine the focusing plane, then the focusing process is simple and fast, but the presence of objects like particles or bubbles causes incorrect focusing planes and reduced measurement precision
Solution Approach 1:
The patent combines gradient image information with original microscopic image information into a fused image that is then processed by a neural network. This merging of multiple image types allows the system to overcome the limitations of using either image type alone, particularly in the presence of distracting objects like particles or bubbles, thereby improving focusing quality measurement precision without requiring overly complex separate processing systems.
Solution Approach 2:
The neural network acts as an intermediary that processes the fused image containing both gradient and original image information. This intermediary component translates complex image data into an accuracy measure of focusing quality, resolving the contradiction by providing a sophisticated yet integrated solution that improves precision while managing complexity through a unified processing architecture.
2Reliability
If only gradient image information is used to determine focusing quality, then the processing is simpler, but the measurement is prone to errors caused by objects like particles or bubbles
Solution Approach 1:
The patent merges gradient image information with original microscopic image information to create a fused image. This combination allows the system to leverage the edge-detection strengths of gradient images while using the contextual information from original images to distinguish actual cellular structures from distracting objects like particles or bubbles, thereby improving reliability without excessive complexity.
Solution Approach 2:
The neural network analyzes multiple parameters from the fused image simultaneously, including but not limited to gradient magnitudes, original image intensities, and their combinations. By changing from single-parameter to multi-parameter analysis, the system achieves more reliable focusing assessment that is resistant to errors caused by particles or bubbles, while the integrated approach manages the complexity of processing multiple parameters.
3Measurement precision
If a neural network processes both gradient and microscopic image information, then the focusing measure accuracy is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent performs gradient calculation and image fusion as preliminary steps before the neural network processing. By preparing the fused image in advance with pre-computed gradient information, the system reduces the computational burden during the actual neural network inference, thereby improving focusing measure accuracy while minimizing additional processing time.
Solution Approach 2:
The neural network processes the fused image by segmenting it into relevant features and patterns that indicate focusing quality. This segmentation approach allows the network to focus computational resources on the most informative aspects of the image data, improving accuracy while reducing overall processing time by avoiding exhaustive analysis of all image pixels.
4Manufacturing precision
If conventional focusing methods are used, then the system is simpler to operate, but the pattern detection accuracy is reduced due to incorrect focusing planes
Solution Approach 1:
The patent combines multiple image processing approaches (gradient analysis, original image analysis, and neural network processing) into a unified system that automatically determines focusing quality. This merging improves pattern detection accuracy by ensuring images are correctly focused before analysis, while the integrated nature of the system manages complexity through a coordinated workflow rather than separate manual steps.
Solution Approach 2:
The neural network provides feedback about the accuracy of focusing by analyzing the fused image and generating an accuracy measure. This feedback mechanism allows the system to automatically adjust or flag images that require refocusing, thereby improving pattern detection accuracy without requiring complex manual intervention, as the feedback loop guides the focusing process efficiently.
Data Source
AI summary
Proposed is a method for determining a focusing measure of a microscopic image, the microscopic image representing an image of a biological cellular substrate, the method comprising: providing the microscopic image, determining a gradient image on the basis of the microscopic image, processing image information from the gradient image and image information from the microscopic image by means of a neural network to determine the focusing measure, the focusing measure indicating quality of focusing in the microscopic image in relation to a cellular substrate plane of the cellular substrate.