Sperm Cell Characterization Using Defocused CNN Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for observing motile microscopic particles, such as sperm cells, face challenges in achieving a balance between spatial resolution, observed field, and depth of field, leading to complex and costly setups, and prolonged computation times when characterizing mobile particles.
Innovation Solution
A method utilizing a detection convolutional neural network to process a series of images acquired with defocused or lensless imaging, estimating particle positions, and applying classification neural networks for morphological and motility characterization without requiring lens displacement or holographic reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional focused microscopy is used to observe motile particles, then spatial resolution is improved, but the observed field size and depth of field are reduced
Solution Approach 1:
The patent changes the focusing parameter by intentionally defocusing the optical system. Instead of maintaining sharp focus at a single plane, the system operates in a defocused state where light from multiple depths contributes to the image, thereby expanding both the observed field and depth of field while retaining sufficient spatial resolution for particle characterization
2Measurement precision
If conventional focused microscopy is used to observe motile particles, then spatial resolution is improved, but depth of field is reduced
Solution Approach 1:
The patent changes the focusing parameter by intentionally defocusing the optical system. Instead of maintaining sharp focus at a single plane, the system operates in a defocused state where light from multiple depths contributes to the image, thereby expanding both the observed field and depth of field while retaining sufficient spatial resolution for particle characterization
3Area of stationary object
If lens displacement is used to compensate for small observed field, then field coverage is improved, but device complexity increases
Solution Approach 1:
The patent extracts and removes the translation plate component from the optical system. By operating in a defocused mode, the system inherently provides a larger observed field without requiring mechanical displacement mechanisms, thereby simplifying the device while maintaining field coverage
Solution Approach 2:
The patent replaces the mechanical lens displacement system with an optical approach. Instead of physically moving the lens to scan different fields, the defocused optical configuration provides wide field coverage inherently, eliminating the need for mechanical translation mechanisms
4Area of stationary object
If lens displacement is used to compensate for small observed field, then field coverage is improved, but accuracy of displacement increases complexity
Solution Approach 1:
The patent extracts and removes the translation plate component from the optical system. By operating in a defocused mode, the system inherently provides a larger observed field without requiring mechanical displacement mechanisms, thereby simplifying the device while maintaining field coverage
Solution Approach 2:
The patent replaces the mechanical lens displacement system with an optical approach. Instead of physically moving the lens to scan different fields, the defocused optical configuration provides wide field coverage inherently, eliminating the need for mechanical translation mechanisms
5Area of stationary object
If holographic reconstruction algorithms are used for lensless imaging, then computation time increases
Solution Approach 1:
The patent extracts and removes the translation plate component from the optical system. By operating in a defocused mode, the system inherently provides a larger observed field without requiring mechanical displacement mechanisms, thereby simplifying the device while maintaining field coverage
Data Source
AI summary
A method for characterizing a mobile particle in a sample includes acquisition of at least one image of the sample during an acquisition period, using an image sensor and formation of a series of images, the series of images comprising at least one image; use of each image of the series of images as input image of a detection convolutional neural network, the detection convolutional neural network being configured to detect the particles and to produce, from each image, an output image on which each detected particle is assigned a distribution of intensity, centered on the particle and extending around the particle; for each detected particle, from each output image, estimation of a position of each detected particle in each image of the series of images; and characterization of each detected particle from the estimation of the position established from each image of the series of images.