Method and system for constructing digital color image depicting sample

By using broadband LEDs and multiple lighting modes in Fourier stacked microscopy, combined with the combined technology of machine learning models, the problem of insufficient imaging efficiency and accuracy in the prior art is solved, and fast and efficient high-resolution color image capture is achieved.

CN119768824APending Publication Date: 2025-04-04CELLAVISION
0 Cites 0 Cited by

Patent Information

Application Number
CN202380061779.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-08-30
Filing Date
2023-08-29
Publication Date
2025-04-04

AI Technical Summary

Technical Problem

Existing digital microscopy techniques have limitations in optical resolution and imaging efficiency, especially when capturing high-resolution color images, requiring multiple illumination and imaging, which is time-consuming and insufficient accuracy.

Method used

Fourier stacked microscopy (FPM) combined with broadband LED lighting, the samples were imaged through multiple illumination modes, multiple digital images were captured, and these images were combined into high-resolution color images using machine learning models.

Benefits of technology

The method of using broadband LEDs in Fourier stacked microscopy is realized, which improves the efficiency and accuracy of sample imaging, can quickly capture high-resolution color images, and reduces time and resource consumption during the imaging process.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The inventive concept relates to a method (30) and an apparatus (10) for training a machine learning model for constructing a digital color image depicting a sample. The method comprises: acquiring (S300) a digital image training set of a training sample by illuminating (S302) the training sample with a plurality of illumination modes by a plurality of white light emitting diodes and capturing (S304) a digital image of the training sample for each of the plurality of illumination modes; receiving (S306) real data comprising a high-resolution digital color image of the training sample, where the high-resolution digital color image has a resolution that is relatively higher than a resolution of at least one digital image in the training set of digital images; and training (S308) the machine learning model for constructing a digital color image of the depicted sample using the digital image training set and the true data. The inventive concept further relates to a microscope system (20) and a method (40) for constructing a digital color image (602) depicting a sample.
Need to check novelty before this filing date? Find Prior Art