Research on Denoising of Cryo-em Images Based on Deep Learning

Abstract

Cryo-em (Cryogenic electron microscopy) is a technology this can build bio-macromolecule of three-dimensional structure. Under the condition of now, the projection image of the biological macromolecule which is collected by the Cryo-em technology that the contrast is low, the signal to noise is low, image blurring, and not easy to distinguish single particle from background, the corresponding processing technology is lagging behind. Therefore, make Cryoem image denoising useful, and maintaining bio-macromolecule of contour or signal of function-construct improve Cryo-em image quality or resolution of Cryo-em three-dimensional structure have important effect. This paper researched a denoising function base on GANs (generative adversarial networks), purpose an improved discriminant model base on Wasserstein distance and an improved image denoising model by add gray constraint. Our model turn discriminant model's training process from binary classification's training process into regression task training process, it make GANs in training process more stable, more reasonable parameter passing. Meantime, we also propose an improved generative model by add gray constraint. The experimental results show that our model can increase the peak signal-to-noise ratio of the Cryo-em simulation image by 10.3 dB and improve SSIM (Structural Similarity Index) of the denoised image results by 0.43. Compared with traditional image denoising algorithms such as BM3D (Block Matching 3D), our model can better save the model structure and the vein signal in the original image and the operation speed is faster.

Generalizable Denoising of Microscopy Images using Generative Adversarial Networks and Contrastive Learning

“…GANs have also been recently used for denoising tasks with success. For example, [25] implement a GAN as a tool to recover structural information from cryoelectron microscopy data; [26] that apply a β-GAN combining GANs and auto-encoders to achieve a robust estimate of certain distributional parameters under Huber contamination model with statistical optimality; or [27], which also use a GAN-based method with a modified discriminator that performs regression and helps to stabilize the training process. Denoising Optical Coherence Tomography is also a task where GANs have been widely used.…”

CLGA Net: Cross Layer Gated Attention Network for Image Dehazing

“…Dehaze-Net [9] processes blurred images by estimating the transmission matrix, but inaccurate transmission mapping estimation reduces the dehazing effect of the model. The idea of applying GAN network to denoising is usually analogic to the use of a generator for generating denoising images, and a discriminator for judging denoising effects [27]. GFN-Net [28] employees a dual-branch [29] convolutional neural network to extract basic features and recovery features respectively.…”

FPD Net: Feature Pyramid DehazeNet

“…For example, Dehaze-Net deal with hazy images by estimating the transmission matrix, but an inaccurate estimation of the transmission map will reduce the model's dehazing effect. The method of using GAN network to denoise often uses a generator to generate a denoised image and a discriminator to judge the effect of denoising, such as [16]. AOD-Net uses a deformed atmospheric scattering model, which generalizes two unknowns in the atmospheric scattering model to a single unknown to reduce the loss in the dehazing process.…”