Random Noise Attenuation In Seismic Data Though Multi-Scale Residual Dense Network
$\to$ 介词不要大写
Random Noise Suppression of Seismic Data through Multi-Scale Residual Dense Network

Random noise suppression is an important issue in seismic data processing, which helps to improve the efficiency and accuracy of seismic data processing and interpretation.
$\to$ 重复的词组
$\to$ 统一词汇
Random noise suppression is an important technique in different seismic data processing stages to improve the efficiency of the process and the accuracy of the interpretation.
Random noise suppression is an important technique to improve the efficiency and accuracy of seismic data processing.

As the neural network has achieved good results in the field of image denoising, many scholars use it to denoise seismic data.
$\to$ 在图像上的效果好，在地震数据上呢？
$\to$ 神经网络是一个宽泛的技术，能不能写得更具体一点？
$\to$ 要描述其优缺点
Convolutional neural network is a popular technique in seismic data denoising, however the results are still not satisfactory.

In this paper, we propose the Random Noise Attenuation In Seismic Data Though Multi-Scale Residual Dense Network (DnSRFD).
$\to$ 提出的是一个方案，一个网络，不需要把数据和问题放进去
In this paper, we propose a Multi-Scale Residual Dense Network (MSRDN) for random noise suppression of seismic raw data.

Firstly, dense blocks closely connected by multi-scale convolutional layers are constructed to enhance the regularization effect of the network.
Secondly, residual blocks composed of different dense blocks are constructed to solve the problem of network degradation.
Finally, the outputs of all residual blocks are combined through deconvolution to obtain DnSRFD.
$\to$ 应该自顶向下描述，从总体结构到具体模块
$\to$ different 表示结构不同？其实并没有。
$\to$ how and why 的节奏
First, the network consists of a shallow feature extraction module, multiple residual modules and a deconvolution layer.
The serve for feature extraction, denoising, and recovering. (自己想下它们各自的作用)
Second, each residual block consists of a number of dense blocks.
They are designed to alleviate network degradation.
Third, dense blocks are closely connected by multi-scale convolutional layers.
They can enhance the regularization effect of the network.

The experimental results of theoretical data show that DnSRFD is more accurate and stable than previous algorithms.