Super-resolution convolutional neural network for the improvement of the image quality of magnified images in chest radiographs

Umehara, Kensuke; Ota, Junko; Ishimaru, Naoki; Ohno, Shunsuke; Okamoto, K.; Suzuki, Takanori; Shirai, Naoki; Ishida, Takayuki

doi:10.1117/12.2249969

Cited by 23 publications

(17 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Later on, these networks started getting used for super-resolution (SR) imaging [93] and some of the first applications focused on photographs (see, e.g., [93][94][95][96]) and movies [97]. They were rapidly used for the resolution enhancement of, for example, satellite images [98,99] and medical images [100], like magnetic resonance imaging [101,102] and CT [103][104][105][106][107]. The success of a lot of these SR applications is explained by the emergence of generative adversarial networks (GANs) [108] which are commonly known for their strength at generating realistic "fake" images [109].…”

mentioning

confidence: 99%

Computed Tomography 3D Super-Resolution with Generative Adversarial Neural Networks: Implications on Unsaturated and Two-Phase Fluid Flow

2020

View full text Add to dashboard Cite

Fluid flow characteristics are important to assess reservoir performance. Unfortunately, laboratory techniques are inadequate to know these characteristics, which is why numerical methods were developed. Such methods often use computed tomography (CT) scans as input but this technique is plagued by a resolution versus sample size trade-off. Therefore, a super-resolution method using generative adversarial neural networks (GANs) was used to artificially improve the resolution. Firstly, the influence of resolution on pore network properties and single-phase, unsaturated, and two-phase flow was analysed to verify that pores and pore throats become larger on average and surface area decreases with worsening resolution. These observations are reflected in increasingly overestimated single-phase permeability, less moisture uptake at lower capillary pressures, and high residual oil fraction after waterflooding. Therefore, the super-resolution GANs were developed which take low (12 µm) resolution input and increase the resolution to 4 µm, which is compared to the expected high-resolution output. These results better predicted pore network properties and fluid flow properties despite the overestimation of porosity. Relevant small pores and pore surfaces are better resolved thus providing better estimates of unsaturated and two-phase flow which can be heavily influenced by flow along pore boundaries and through smaller pores. This study presents the second case in which GANs were applied to a super-resolution problem on geological materials, but it is the first one to apply it directly on raw CT images and to determine the actual impact of a super-resolution method on fluid predictions.Materials 2020, 13, 1397 2 of 33 examples requires detailed knowledge on fluid storage and flow properties which could be estimated through experiments, either carried out in specialized laboratories or deduced from digital rocks. Laboratory measurements, however, require expensive and complex set-ups, take a long time to complete, are often destructive, do not always entirely capture the studied properties and on top of that, experiments often fail to produce reproducible results between laboratories [9,[12][13][14][15][16][17][18][19]. Therefore, numerical solutions have been developed allowing to compute fluid storage and flow properties on digital rocks [7,[20][21][22][23]. Such digital rocks are acquired through direct imaging or numerical simulations of rocks.Computed tomography (CT) is a technique to rapidly capture the three-dimensional structure of materials [22,24,25]. However, like any imaging technique, CT is plagued by a sample size versus resolution trade-off, which is adversely affecting fluid flow simulations. Both a high resolution and sufficiently large volume are required for accurate fluid flow estimations at the pore-scale [9,[26][27][28]. Resolution impacts how well pore network topology and pore surfaces are resolved. At lower resolution, i.e., lower resolving power, pore sizes are often overestimated and specific surf...

show abstract

mentioning

confidence: 99%

Computed Tomography 3D Super-Resolution with Generative Adversarial Neural Networks: Implications on Unsaturated and Two-Phase Fluid Flow

2020

View full text Add to dashboard Cite

show abstract

“…The method proposed in this paper is compared with several state-of-art SR methods, including computer vision and remote sensing SR methods, i.e., Bicubic, SRCNN [47], VDSR [34], and 3D-FCNN [23]. SRCNN, VDSR, and 3D-FCNN followed the default settings as described in [23], [34], [47]. In order to ensure the fairness of the experiment, the three neural networks and FS-3DCNN are trained on the same datasets, and stopped until the indicator of the validating set does not improve.…”

Section: B Compared With the State-of-art Methodsmentioning

confidence: 99%

A Frequency-Separated 3D-CNN for Hyperspectral Image Super-Resolution

Wang

Shi

2020

IEEE Access

View full text Add to dashboard Cite

Considering the limitations such as cost, it is of great significance to use super-resolution methods to improve image spatial quality in the field of hyperspectral remote sensing. Due to little dependence on auxiliary information which is difficult to obtain, i.e., multispectral images and natural images, methods based on single-frame are generally considered to have good flexibility and application value. In this paper, a three-dimensional convolutional neural network with three branches combined with an analytical method is proposed, achieving better SR quality and suppressing spectral distortion as well. Firstly, the wavelet transformation is introduced to decompose the hyperspectral image into a variety frequency of components effectively and reversibly. Then, these components are fed into different three-dimensional convolutional branches respectively. Finally, hyperspectral images with high resolution are obtained by dimension amplification, detail reconstruction and inverse wavelet transformation. The presence of frequency separation and the architecture of our model having different branches designed according to frequency make it better than comparable approaches. The method proposed in this paper not only combines the high efficiency of analytical method and the flexibility of neural network, but inhibits the influence of spectral distortion as well. Compared with the state-of-art methods on real space-based hyperspectral image datasets, the effectiveness of the proposed method is demonstrated. INDEX TERMS Hyperspectral remote sensing, super-resolution, wavelet transformation, neural network. LIGUO WANG (Member, IEEE) received the M.S. and Ph.D. degrees in signal and information processing from the

show abstract

“…SRCNN learns an end-to-end mapping from low-to high-resolution images. In [6,7], the authors applied and evaluated the SRCNN method to improve the image quality of magnified images in chest radiographs and CT images. Moreover, [9] introduced an efficient sub-pixel convolution network (ESPCN), which was shown to be more computationally efficient than SRCNN.…”

Section: Introductionmentioning

confidence: 99%

Computed Tomography Image Enhancement Using 3D Convolutional Neural Network

Shen

Gao

et al. 2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Computed tomography (CT) is increasingly being used for cancer screening, such as early detection of lung cancer. However, CT studies have varying pixel spacing due to differences in acquisition parameters. Thick slice CTs have lower resolution, hindering tasks such as nodule characterization during computer-aided detection due to partial volume effect. In this study, we propose a novel 3D enhancement convolutional neural network (3DECNN) to improve the spatial resolution of CT studies that were acquired using lower resolution/slice thicknesses to higher resolutions. Using a subset of the LIDC dataset consisting of 20,672 CT slices from 100 scans, we simulated lower resolution/thick section scans then attempted to reconstruct the original images using our 3DECNN network. A significant improvement in PSNR (29.3087dB vs. 28.8769dB, p-value < 2.2e − 16) and SSIM (0.8529dB vs. 0.8449dB, pvalue < 2.2e − 16) compared to other state-of-art deep learning methods is observed.

show abstract

Super-resolution convolutional neural network for the improvement of the image quality of magnified images in chest radiographs

Cited by 23 publications

References 7 publications

Computed Tomography 3D Super-Resolution with Generative Adversarial Neural Networks: Implications on Unsaturated and Two-Phase Fluid Flow

Computed Tomography 3D Super-Resolution with Generative Adversarial Neural Networks: Implications on Unsaturated and Two-Phase Fluid Flow

A Frequency-Separated 3D-CNN for Hyperspectral Image Super-Resolution

Computed Tomography Image Enhancement Using 3D Convolutional Neural Network

Contact Info

Product

Resources

About