Construction Method of Microscope Deconvolutional Neural Network Model Based on Richardson-Lucy Algorithm

Abstract

The invention discloses a method for constructing a microscope deconvolution neural network model based on the Richardson-Lucy algorithm. A large number of fluorescence microscope images in a training set and corresponding deconvoluted samples are estimated to be input into a neural network constructed. A deconvolution relationship between acquired images blurred by fluorescence microscopy and sample estimates as labels is learned. It is worth mentioning that this network is universal. For the data collected by the same fluorescence microscope, it has shown good deblurring ability during the test process, and it also has good generalization for different types of data sets. ability. In general, the present invention uses a neural network based on the Richardson-Lucy deconvolution structure to build a universal framework, and establishes a deconvolution mapping relationship through powerful feature extraction capabilities, thereby completing Non-iterative methods for deblurring fluorescence microscopy images.

Description

technical field [0001] The invention belongs to the technical field of fluorescence microscopic imaging, and in particular relates to a method for constructing a microscope deconvolution neural network model based on the Richardson-Lucy algorithm. Background technique [0002] Optical microscopy has a long history and is one of the most powerful means in modern biological research. The development of fluorescent labeling technology has promoted the development of microscopy technology towards higher resolution and higher contrast. However, due to the diffraction of light, the microscopic images are inherently blurred; for 3D fluorescence microscopic images, since only a part of the light is collected by the microscope from one direction, there is also severe resolution anisotropy, and the limited resolution is usually not sufficient. Biological research requires visualization of samples and can cause difficulties in post-processing of fluorescence microscopy images. The poi...

AI Technical Summary

Problems solved by technology

Ideally, researchers hope to automatically find a suitable model from sample data to solve a certain type of problem through a data-driven method. The Chinese patent with the publication number CN111524078A proposes a dense network-based microscope image deblurring method. From the perspective of neural network, the deconvolution problem of the same fluorescence microscopy data is solved. However, this method is similar to a black box, lacks reasonable interpretation, sometimes introduces false artifacts, and also limits the generalization of the network. ability

Images