Deep Learning-Based Super-Resolution Applied to Dental Computed Tomography

Hatvani, Janka; Horváth, András; Michetti, Jérôme; Basarab, Adrian; Kouamé, Denis; Gyöngy, Miklós

doi:10.1109/trpms.2018.2827239

Cited by 87 publications

(58 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…While parametric modelling approaches have been applied in order to enhance medical imaging [8], recent work using deep learning approaches has also shown the potential for the application of neural networks (NNs) for enhancing image resolution. Examples for such an application include, a 4x upscaling on photographic images [9], optical microscopy (improving the resolution from 40x to 100x) [10], dental imaging [11], phase imaging [12], fluorescence microscopy [13], magnetic resonance imaging [14], SEM imaging [15,16], positronemission tomography [17], stochastic optical reconstruction microscopy [18], and ultrasound imaging [19]. NNs are also well-suited to the classification of objects in images, and accordingly the classification of biological, pollution and colloidal particles from images and scattering patterns has also been demonstration [20][21][22][23][24].…”

Section: Introductionmentioning

confidence: 99%

A neural lens for super-resolution biological imaging

et al. 2019

View full text Add to dashboard Cite

Visualizing structures smaller than the eye can see has been a driving force in scientific research since the invention of the optical microscope. Here, we use a network of neural networks to create a neural lens that has the ability to transform 20× optical microscope images into a resolution comparable to a 1500× scanning electron microscope image. In addition to magnification, the neural lens simultaneously identifies the types of objects present, and hence can label, colour-enhance and remove specific types of objects in the magnified image. The neural lens was used for the imaging of Iva xanthiifolia and Galanthus pollen grains, showing the potential for low cost, non-destructive, highresolution microscopy with automatic image processing.

show abstract

Section: Introductionmentioning

confidence: 99%

A neural lens for super-resolution biological imaging

et al. 2019

View full text Add to dashboard Cite

show abstract

“…In our first ablation experiment, we have eliminated all the enhancements in order to show the performance level of a baseline CNN on the CH data set. Since there are several SISR works [3], [6], [16], [18] based on the standard ESPCN model [19], we have eliminated the second convolutional block in the second ablation experiment, transforming our architecture into a standard ESPCN architecture. The performance drops from 0.9270 to 0.9236 in terms of SSIM and from 36.22 to 35.94 in terms of PSNR.…”

Section: Ablation Study Resultsmentioning

confidence: 99%

“…The first CNN increases the resolution on two axes (width and height), while the second CNN takes the output from the first CNN and further increases the resolution on the third axis (depth). Different from related methods [3], [6], [16], [18], we compute the loss with respect to the groundtruth high-resolution image right after the upscaling layer, in addition to computing the loss after the last convolutional layer. The intermediate loss forces our network to produce a better output, closer to the ground-truth.…”

Section: Introductionmentioning

confidence: 99%

Convolutional Neural Networks With Intermediate Loss for 3D Super-Resolution of CT and MRI Scans

Ionescu

Verga

2020

IEEE Access

View full text Add to dashboard Cite

Computed Tomography (CT) scanners that are commonly-used in hospitals and medical centers nowadays produce low-resolution images, e.g. one voxel in the image corresponds to at most onecubic millimeter of tissue. In order to accurately segment tumors and make treatment plans, radiologists and oncologists need CT scans of higher resolution. The same problem appears in Magnetic Resonance Imaging (MRI). In this paper, we propose an approach for the single-image super-resolution of 3D CT or MRI scans. Our method is based on deep convolutional neural networks (CNNs) composed of 10 convolutional layers and an intermediate upscaling layer that is placed after the first 6 convolutional layers. Our first CNN, which increases the resolution on two axes (width and height), is followed by a second CNN, which increases the resolution on the third axis (depth). Different from other methods, we compute the loss with respect to the ground-truth high-resolution image right after the upscaling layer, in addition to computing the loss after the last convolutional layer. The intermediate loss forces our network to produce a better output, closer to the ground-truth. A widely-used approach to obtain sharp results is to add Gaussian blur using a fixed standard deviation. In order to avoid overfitting to a fixed standard deviation, we apply Gaussian smoothing with various standard deviations, unlike other approaches. We evaluate the proposed method in the context of 2D and 3D super-resolution of CT and MRI scans from two databases, comparing it to related works from the literature and baselines based on various interpolation schemes, using 2× and 4× scaling factors. The empirical study shows that our approach attains superior results to all other methods. Moreover, our subjective image quality assessment by human observers reveals that both doctors and regular annotators chose our method in favor of Lanczos interpolation in 97.55% cases for an upscaling factor of 2× and in 96.69% cases for an upscaling factor of 4×. In order to allow others to reproduce our state-of-the-art results, we provide our code as open source at https://github.com/lilygeorgescu/3d-super-res-cnn. INDEX TERMSConvolutional neural networks, single-image super-resolution, CT images, MRI images, medical image super-resolution.

show abstract

“…The high-frequency band was suppressed by a Hanning-window before computing the inverse Fouriertransform and averaging the 13 estimated PSFs. For further details see [16]. Finally a 3D Gaussian function was fitted to the averaged PSF to estimate the standard deviations σ 1 , σ 2 and σ 3 .…”

Section: Psf Estimationmentioning

confidence: 99%

“…Lowrank [10], or wavelet representations [11] have also proved to be efficient tools for SR. A method based on a sparse representation was applied to 3D MRI images with a patchbased structural similarity constraint in [12]. Convolutional neural networks have also shown interesting properties for SR, where the network is trained to map an LR image to its HR counterpart [13]- [16]. However, this technique requires a large training dataset, which is not always available.…”

mentioning

confidence: 99%

A Tensor Factorization Method for 3-D Super Resolution With Application to Dental CT

Hatvani

Basarab

Tourneret

et al. 2019

IEEE Trans. Med. Imaging

Self Cite

View full text Add to dashboard Cite

Available super-resolution techniques for 3D images are either computationally inefficient prior-knowledge-based iterative techniques or deep learning methods which require a large database of known low-and high-resolution image pairs. A recently introduced tensor-factorization-based approach offers a fast solution without the use of known image pairs or strict prior assumptions. In this article this factorization framework is investigated for single image resolution enhancement with an off-line estimate of the system point spread function. The technique is applied to 3D cone beam computed tomography for dental image resolution enhancement. To demonstrate the efficiency of our method, it is compared to a recent state-ofthe-art iterative technique using low-rank and total variation regularizations. In contrast to this comparative technique, the proposed reconstruction technique gives a 2-order-of-magnitude improvement in running time -2 minutes compared to 2 hours for a dental volume of 282×266×392 voxels. Furthermore, it also offers slightly improved quantitative results (peak signalto-noise ratio, segmentation quality). Another advantage of the presented technique is the low number of hyperparameters. As demonstrated in this paper, the framework is not sensitive to small changes of its parameters, proposing an ease of use.

show abstract

Deep Learning-Based Super-Resolution Applied to Dental Computed Tomography

Abstract: Learning-Based Super-Resolution Applied to Dental Computed Tomography. (2019) IEEE Transactions on Radiation and Plasma Medical Sciences, 3 (2). 120-128.

Cited by 87 publications

References 42 publications

A neural lens for super-resolution biological imaging

A neural lens for super-resolution biological imaging

Convolutional Neural Networks With Intermediate Loss for 3D Super-Resolution of CT and MRI Scans

A Tensor Factorization Method for 3-D Super Resolution With Application to Dental CT

Contact Info

Product

Resources

About