3D Image Matching for Discrete Biological Entities Using Marker Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to efficiently track and differentiate large numbers of embedded biological samples, particularly rare cells, in 3D cell cultures, making it difficult to identify and isolate individual entities within a collective sample.
Innovation Solution
A method and imaging system that generate and match three-dimensional images of discrete entities by determining vectors from reference items to constituent parts of a marker, using properties like vector length and angle, and generating rotation-invariant representations for efficient comparison and matching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If large numbers of cells are handled in a single vessel, then the ability to analyze rare cells is improved, but the ability to track individual entities is lost
Solution Approach 1:
The system segments the large cell population into individually trackable entities by assigning unique identification codes to each discrete entity (hydrogel bead). This segmentation allows rare cells to be analyzed within large populations while maintaining the ability to track individual entities through their unique codes, resolving the contradiction between handling large quantities and preserving tracking information.
Solution Approach 2:
The patent introduces an intermediary identification system consisting of markers with unique codes on discrete entities. These markers act as intermediaries between the cells and the imaging system, enabling the system to track individual entities within large populations by detecting and recording the unique codes assigned to each entity.
2Reliability
If 3D cell culture with discrete entities is used, then cultivation of rare cells is improved, but image matching and tracking between time points becomes difficult
Solution Approach 1:
The system performs preliminary action by assigning unique identification codes to discrete entities before imaging. This pre-coding allows for efficient matching of images taken at different time points, as the system can directly compare the unique codes rather than attempting to match complex 3D structures, thereby reducing the complexity of the image matching system while maintaining reliable cultivation tracking.
Solution Approach 2:
The patent uses optical copying of the unique identification codes from the discrete entities into digital form. The imaging system captures images of the markers and extracts their unique codes, creating a digital copy that can be easily stored, compared, and matched across different time points, simplifying the tracking process compared to analyzing raw 3D images directly.
3Measurement precision
If unique identification of individual entities is implemented, then tracking capability is improved, but the complexity of generating rotation-invariant representations increases
Solution Approach 1:
Instead of attempting to recognize and match complex 3D structures of discrete entities, the system inverts the approach by placing simple, detectable markers with unique codes onto the entities. The imaging system then detects these simplified markers rather than analyzing the complex entity structures, achieving high identification precision while reducing the complexity of the representation generation system.
Solution Approach 2:
The patent replaces complex mechanical/image processing operations with optical detection of fluorescent markers. Rather than using complex algorithms to recognize and match 3D structures, the system uses optical detection to read the unique codes on markers, substituting a simpler optical measurement system for a complex image analysis system.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method is provided for matching images (103) of at least one discrete entity (100) to each other, the discrete entity (100) comprising a biological sample (108) and a plurality of constituent parts of a marker (102), the method comprising the following steps: generating a first representation (200, 302, 304, 306) of the marker (102) from the three-dimensional first image; generating a second representation of the marker (102) from the three-dimensional second image; and matching the three-dimensional first image with the three-dimensional second image, when the first representation and the second representation match, or rejecting the match, when the first representation and the second representation do not match. The steps for generating the representations comprise: determining vectors (104) from at least one reference item (112) to at least some of the constituent parts (110) of the marker (102), determining for the vectors (104) at least one value of a property of the vectors (104), and generating the representation of the marker based on the frequency of the values of the property. In a further aspect an imaging device is provided for matching images (103) of at least one discrete entity (100) to each other.