Image denoising via deep residual convolutional neural networks

Lan, Rushi; Zou, Haizhang; Pang, Cheng; Zhong, Yanru; Liu, Zhenbing; Luo, Xiangfeng

doi:10.1007/s11760-019-01537-x

Cited by 38 publications

(13 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Changes in the closing price on the forecast date reveal how much the price rose or fell (after 1 day or 5 days) based on the last day of the 30 days after the image was created (see Eqs (1), ( 2)). If the image name rate of return was greater than 0, 1 was automatically added to the (3,6,12), (4,8,16), (5,10,20), (6,12,24), (7,14,28), (8,16,32), (9,18,36) (10,20,40), (11,22,44), (12,24,48), (13,26,52), (14,28,56), (15,30,60), (16,32,64) https://doi.org/10.1371/journal.pone.0253121.t003…”

Section: Datamentioning

confidence: 99%

“…In accordance with previous studies [48,50,51], among the possible convolution layers, pooling layers, dropouts, and filters, we focused on the number of filters and dropouts, examining the activation function for each layer as the hyperparameter to improve prediction. To determine whether the shape of dropouts and filters affects the accuracy of CNN inferences, the dropout variable was set to (0.25,0.5) and (0.5,0.5), and the number of 3-layer convolution kernels was set to (1,2,4), (2,4,8), (3,6,12), (4,8,16), (5,10,20), (6,12,24), (7,14,28), (8,16,32), (9,18,36), (10,20,40), (11,22,44), (12,24,48), (13,26,52), (14,28,56), (15,30,…”

Section: Plos Onementioning

confidence: 99%

“…However, most stock price chart types are fixed so that the characteristics of the chart image that can affect stock price prediction are not easily understood [2,12]. In deep learning algorithms such as CNNs, since the characteristics of the image affect predictive performance, the shape of the chart may be an important variable [11,13]. However, only a few studies examine how the characteristics of stock price charts affect the performance of deep learning algorithms (especially CNNs) for stock price prediction.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Impact of chart image characteristics on stock price prediction with a convolutional neural network

Jin

Kwon

2021

PLoS ONE

View full text Add to dashboard Cite

Stock price prediction has long been the subject of research because of the importance of accuracy of prediction and the difficulty in forecasting. Traditionally, forecasting has involved linear models such as AR and MR or nonlinear models such as ANNs using standardized numerical data such as corporate financial data and stock price data. Due to the difficulty of securing a sufficient variety of data, researchers have recently begun using convolutional neural networks (CNNs) with stock price graph images only. However, we know little about which characteristics of stock charts affect the accuracy of predictions and to what extent. The purpose of this study is to analyze the effects of stock chart characteristics on stock price prediction via CNNs. To this end, we define the image characteristics of stock charts and identify significant differences in prediction performance for each characteristic. The results reveal that the accuracy of prediction is improved by utilizing solid lines, color, and a single image without axis marks. Based on these findings, we describe the implications of making predictions only with images, which are unstructured data, without using large amounts of standardized data. Finally, we identify issues for future research.

show abstract

Section: Datamentioning

confidence: 99%

Section: Plos Onementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Impact of chart image characteristics on stock price prediction with a convolutional neural network

Jin

Kwon

2021

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…Convolutional neural network (CNN) is a trainable feedforward network [17]. CNN has the characteristics of representation learning, so it maintains translation invariance and scaling invariance for input information to a certain extent.…”

Section: Convolutional Neural Networkmentioning

confidence: 99%

Image fusion algorithm in Integrated Space-Ground-Sea Wireless Networks of B5G

Cui

Wang

et al. 2021

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

In recent years, in Space-Ground-Sea Wireless Networks, the rapid development of image recognition also promotes the development of images fusion. For example, the content of a single-mode medical image is very single, and the fused image contains more image information, which provides a more reliable basis for diagnosis. However, in wireless communication and medical image processing, the image fusion effect is poor and the efficiency is low. To solve this problem, an image fusion algorithm based on fast finite shear wave transform and convolutional neural network is proposed for wireless communication in this paper. This algorithm adopts the methods such as fast finite shear wave transform (FFST), reducing the dimension of the convolution layer, and the inverse process of fast finite shear wave transform. The experimental results show that the algorithm has a very good effect in both objective indicators and subjective vision, and it is also very feasible in wireless communication.

show abstract

“…A group researcher used realtime image noise filter to reduce the noise of MRI images and found that they could improve the signal-to-noise ratio by 60%, without introducing image artifacts and prolonging the detection time [2]. In other study, they used the new deep residual convolutional neural network for image denoising, and the results showed that its denoising effect was better than other algorithms [3]. Some of researchers proposed a pulse-coupled neural network with multi-channel link and feed field, which can be used in satellite image segmentation to improve the processing speed [4].…”

Section: Introductionmentioning

confidence: 99%

Effect of Functional Motion Image Analysis Under Multi-Layer Convolutional Neural Network on Improving Speed and Force

Long¹,

Thomas²

2020

Malays. sport. j.

View full text Add to dashboard Cite

To explore the effect of applying multi-layer convolutional neural network to noise removal during the acquisition and transmission of functional motion images, the classical image denoising algorithms of mean filtering, median filtering, and wavelet transform filtering are first introduced. In addition, the evaluation methods of mean square error (MSE), image enhancement factor (IEF), peak signal to noise ratio (PSNR) and structural similarity measure (SSIM) in the image quality evaluation index system are introduced. Based on the convolutional neural network model, a multi-scale parallel convolutional neural network (MP-CNN) model is constructed to remove the noise in the image, and the functional action image is devoted to different degrees. Finally, the denoising effect is evaluated by the objective and subjective evaluation system. The objective evaluation results show that MP-CNN's MSE, IEF, PSNR, and SSIM are better than the single-channel model, and the test time is shorter. The subjective evaluation results show that the MP-CNN model has the best effect on noise removal after 25 denoising of functional action images. In this study, a multi-channel image denoising model based on the multi-layer convolutional neural network can improve the effect of functional motion image noise removal. KEYWORDSMulti-layer convolutional neural network, Functional action-image, Noise, Evaluation system METHODOLOGY Image denoising algorithmNoise generally has the characteristics of random distribution, and the probability density function can be used to describe the noise distribution, which can be divided into Gaussian, Rayleigh, exponent, γ, uniform and impulse noise and so on. The classical denoising methods include mean filtering, median filtering and wavelet transform filtering. Mean filtering is a common spatial linear filtering method, which plays an obvious role in smoothing noise, and the Box template is often used to filter noise images.

show abstract

Image denoising via deep residual convolutional neural networks

Cited by 38 publications

References 26 publications

Impact of chart image characteristics on stock price prediction with a convolutional neural network

Impact of chart image characteristics on stock price prediction with a convolutional neural network

Image fusion algorithm in Integrated Space-Ground-Sea Wireless Networks of B5G

Effect of Functional Motion Image Analysis Under Multi-Layer Convolutional Neural Network on Improving Speed and Force

Contact Info

Product

Resources

About