Light field microscopic three-dimensional reconstruction method and device based on deep learning algorithm

Abstract

The invention discloses a light field microscopic three-dimensional reconstruction method and device based on a deep learning algorithm, and the method comprises the steps: building a light field microscopic imaging system; simulating the light field microscopic imaging system to obtain a point spread function of the imaging system; generating a simulation sample body distribution data set; generating a simulation light field data set through a point spread function and corresponding sample data, and correcting possibly generated noise and a background model; constructing a deep convolutionalneural network, taking the simulation light field data as the input of the network, taking the simulation sample body distribution data as the output of the network, and training the network accordingto a microscopic sample design loss function; and after the training of the model is completed, inputting the light field microscopic data to be reconstructed into the model for testing, and obtaining a prediction value of the distribution data of the corresponding sample body. According to the method, rapid, high-resolution and less-artifact light field data three-dimensional reconstruction canbe realized while the advantage of rapid acquisition of three-dimensional information by the light field is maintained.

Description

technical field [0001] The present invention relates to technical fields such as computational optics, computational photography, computer vision and computer graphics, and in particular to a light field microscopic three-dimensional reconstruction method and device based on a deep learning algorithm. Background technique [0002] The development of life science and medical technology has put forward higher requirements for 3D fast volume imaging technology. In this case, light field microscopy has become a method that has attracted much attention due to its characteristics of obtaining both spatial and angular information of the sample. Since the light field can quickly obtain the three-dimensional information of the sample, it has become a general solution to the fast imaging problem. After being introduced into optical microscopy, light-field microscopy plays an increasingly important role in bio-optical imaging problems such as three-dimensional imaging problems such as...

AI Technical Summary

Problems solved by technology

[0003] Although light field microscopy technology has the ability to quickly obtain three-dimensional information of samples, there are some inherent problems in its commonly used reconstruction algorithm - three-dimensional deconvolution algorithm, which seriously limits the expansion of the application range of light field microscopy.

First, in order to achieve accurate results, the fitting method needs to have a more accurate estimate of the point spread function, but this is difficult to measure in experiments

Second, the 3D deconvolution algorithm based on maximum likelihood estimation often requires a large number of iterative steps to achieve a better convergence effect, which limits the reconstruction speed

Third, due to the physical limitation of the resolution at the focal plane, traditional methods cannot effectively solve this problem

In addition, due to the ill-conditioned nature of the problem, 3D deconvolution algorithms often cause unpredictable noise

In summary, it can be found that although light field microscopy has significant advantages in some aspects compared with other 3D imaging methods, its reconstruction algorithm still needs to be improved.

Images