Micro-object Detection in Microfluidic Devices Using Pixel Masks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Efficient and robust detection of micro-objects, such as biological cells or beads, in microfluidic environments is challenging due to translucent appearances and non-uniform backgrounds, especially in images like fluorescent images where micro-objects are not illuminated.
Innovation Solution
The method involves generating pixel masks from illuminated images using machine learning algorithms like convolutional neural networks to identify micro-objects, and then using these masks to detect and characterize micro-objects in corresponding non-illuminated images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional image processing methods are used to detect micro-objects in fluorescent images, then the detection process is simple, but the detection reliability is poor due to translucent appearance and non-uniform backgrounds
Solution Approach 1:
The system performs preliminary detection and localization of micro-objects in illuminated images (bright field) before analyzing non-illuminated images (fluorescent). The pixel mask generation from illuminated images provides prior information about micro-object positions and boundaries, which is then used to guide the detection in fluorescent images, improving reliability without requiring complex standalone fluorescent detection algorithms
Solution Approach 2:
The patent introduces pixel masks as an intermediary representation that bridges illuminated and non-illuminated images. The pixel masks encode micro-object characteristics from illuminated images and serve as a mediator to improve detection in fluorescent images where micro-objects are not directly visible, allowing the system to leverage information from both image types without directly combining them
2Measurement precision
If machine learning algorithms are used to generate pixel masks from illuminated images, then the detection precision is improved, but the processing time increases
Solution Approach 1:
The system applies machine learning algorithms selectively to generate pixel masks only for regions containing micro-objects in illuminated images, rather than processing entire images. The convolutional neural network focuses computational resources on identifying and characterizing micro-object boundaries and features, achieving high precision while reducing overall processing time by avoiding unnecessary computation in background regions
Data Source
AI summary
Methods are provided for the automated detection, characterization, and selection of micro-objects in a microfluidic device. In addition, methods are provided for grouping detected micro-objects into subgroups that share the same characteristics and, optionally, repositioning micro-objects in a selected sub-population within the microfluidic device. For example, micro-objects in a selected sub-population can be moved into sequestration pens. The methods also provide for visual displays of the micro-object characteristics, such as two- or three-dimensional graphs, and for user-based definition and/or selection of sub-populations of the detected micro-objects. In addition, non-transitory computer-readable medium in which a program is stored and systems for carrying out any of the disclosed methods are provided.