Methods for image denoising using convolutional neural network: a review

Denoising and Baseline Correction Methods for Raman Spectroscopy Based on Convolutional Autoencoder: A Unified Solution

Preprocessing plays a key role in Raman spectral analysis. However, classical preprocessing algorithms often have issues with reducing Raman peak intensities and changing the peak shape when processing spectra. This paper introduces a unified solution for preprocessing based on a convolutional autoencoder to enhance Raman spectroscopy data. One is a denoising algorithm that uses a convolutional denoising autoencoder (CDAE model), and the other is a baseline correction algorithm based on a convolutional autoencoder (CAE+ model). The CDAE model incorporates two additional convolutional layers in its bottleneck layer for enhanced noise reduction. The CAE+ model not only adds convolutional layers at the bottleneck but also includes a comparison function after the decoding for effective baseline correction. The proposed models were validated using both simulated spectra and experimental spectra measured with a Raman spectrometer system. Comparing their performance with that of traditional signal processing techniques, the results of the CDAE-CAE+ model show improvements in noise reduction and Raman peak preservation.

An efficient lightweight network for image denoising using progressive residual and convolutional attention feature fusion

While deep learning has become the go-to method for image denoising due to its impressive noise removal capabilities, excessive network depth often plagues existing approaches, leading to significant computational burdens. To address this critical bottleneck, we propose a novel lightweight progressive residual and attention mechanism fusion network that effectively alleviates these limitations. This architecture tackles both Gaussian and real-world image noise with exceptional efficacy. Initiated through dense blocks (DB) tasked with discerning the noise distribution, this approach substantially reduces network parameters while comprehensively extracting local image features. The network then adopts a progressive strategy, whereby shallow convolutional features are incrementally integrated with deeper features, establishing a residual fusion framework adept at extracting encompassing global features relevant to noise characteristics. The process concludes by integrating the output feature maps from each DB and the robust edge features from the convolutional attention feature fusion module (CAFFM). These combined elements are then directed to the reconstruction layer, ultimately producing the final denoised image. Empirical analyses conducted in environments characterized by Gaussian white noise and natural noise, spanning noise levels 15-50, indicate a marked enhancement in performance. This assertion is quantitatively corroborated by increased average values in metrics such as Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Feature Similarity Index for Color images (FSIMc), outperforming the outcomes of more than 20 existing methods across six varied datasets. Collectively, the network delineated in this research exhibits exceptional adeptness in image denoising. Simultaneously, it adeptly preserves essential image features such as edges and textures, thereby signifying a notable progression in the domain of image processing. The proposed model finds applicability in a range of image-centric domains, encompassing image processing, computer vision, video analysis, and pattern recognition.

A Review of deep learning methods for denoising of medical low-dose CT images

To prevent patients from being exposed to excess of radiation in CT imaging, the most common solution is to decrease the radiation dose by reducing the X-ray, and thus the quality of the resulting low-dose CT images (LDCT) is degraded, as evidenced by more noise and streaking artifacts. Therefore, it is important to maintain high quality CT image while effectively reducing radiation dose. In recent years, with the rapid development of deep learning technology, deep learning-based LDCT denoising methods have become quite popular because of their data-driven and high-performance features to achieve excellent denoising results. However, to our knowledge, no relevant article has so far comprehensively introduced and reviewed advanced deep learning denoising methods such as Transformer structures in LDCT denoising tasks. Therefore, based on the literatures related to LDCT image denoising published from year 2016-2023, and in particular from 2020 to 2023, this study presents a systematic survey of current situation, and challenges and future research directions in LDCT image denoising field. Four types of denoising networks are classified according to the network structure: CNN-based, Encoder-Decoder-based, GAN-based, and Transformer-based denoising networks, and each type of denoising network is described and summarized from the perspectives of structural features and denoising performances. Representative deep-learning denoising methods for LDCT are experimentally compared and analyzed. The study results show that CNN-based denoising methods capture image details efficiently through multi-level convolution operation, demonstrating superior denoising effects and adaptivity. Encoder-decoder networks with MSE loss, achieve outstanding results in objective metrics. GANs based methods, employing innovative generators and discriminators, obtain denoised images that exhibit perceptually a closeness to NDCT. Transformer-based methods have potential for improving denoising performances due to their powerful capability in capturing global information. Challenges and opportunities for deep learning based LDCT denoising are analyzed, and future directions are also presented.

High-resolution temporal gravity field data products: Monthly mass grids and spherical harmonics from 1994 to 2021

Since April 2002, Gravity Recovery and Climate Experiment (GRACE) and GRACE-FO (FollowOn) satellite gravimetry missions have provided precious data for monitoring mass variations within the hydrosphere, cryosphere, and oceans with unprecedented accuracy and resolution. However, the long-term products of mass variations prior to GRACE-era may allow for a better understanding of spatio-temporal changes in climate-induced geophysical phenomena, e.g., terrestrial water cycle, ice sheet and glacier mass balance, sea level change and ocean bottom pressure (OBP). Here, climate-driven mass anomalies are simulated globally at 1.0° × 1.0° spatial and monthly temporal resolutions from January 1994 to January 2021 using an in-house developed hybrid Deep Learning architecture considering GRACE/-FO mascon and SLR-inferred gravimetry, ECMWF Reanalysis-5 data, and normalized time tag information as training datasets. Internally, we consider mathematical metrics such as RMSE, NSE and comparisons to previous studies, and externally, we compare our simulations to GRACE-independent datasets such as El-Nino and La-Nina indexes, Global Mean Sea Level, Earth Orientation Parameters-derived low-degree spherical harmonic coefficients, and in-situ OBP measurements for validation.

Image denoising and segmentation model construction based on IWOA-PCNN

The research suggests a method to improve the present pulse coupled neural network (PCNN), which has a complex structure and unsatisfactory performance in image denoising and image segmentation. Then, a multi strategy collaborative improvement whale optimization algorithm (WOA) is proposed, and an improved whale optimization algorithm (IWOA) is constructed. IWOA is used to find the optimal parameter values of PCNN to optimize PCNN. By combining the aforementioned components, the IWOA-PCNN model had the best image denoising performance, and the produced images were crisper and preserve more information. IWOA-PCNN processed pictures have an average PSNR of 35.87 and an average MSE of 0.24. The average processing time for photos with noise is typically 24.80 s, which is 7.30 s and 7.76 s faster than the WTGAN and IGA-NLM models, respectively. Additionally, the average NU value measures 0.947, and the average D value exceeds 1000. The aforementioned findings demonstrate that the suggested method can successfully enhance the PCNN, improving its capability for image denoising and image segmentation. This can, in part, encourage the use and advancement of the PCNN.

Overcoming Periodic Stripe Noise in Infrared Linear Array Images: The Fourier-Assisted Correlative Denoising Method

Infrared linear array detectors frequently experience vertical, low-frequency, and periodic stripe noise during imaging, stemming from electro-mechanical interference. Unlike conventional periodic disturbances, this interference showcases long periodicities and is uniquely columnar in orientation. Its presence, especially within the low-frequency domain, renders conventional filtering techniques ineffective and, at times, detrimental to image quality. Addressing this challenge, we introduce Fourier-Assisted Correlative Denoising (FACD), a correlation-centric denoising approach tailored for such unique interference patterns. This mechanism begins with the capture of a pure background image, inclusive of periodic noise, during the non-uniform correction phase of the infrared detector. Leveraging the noise's frequency domain attributes, we extract a one-dimensional single-cycle noise signal. The infrared image is subsequently segmented into parts, and using the detected noise periodicity, the one-dimensional signals for each segment are computed. By leveraging the correlation between these signals and the benchmark one-dimensional noise pattern, we ascertain the noise profile within each segment. This profile is then employed for spatial domain denoising across the entire image frame. Empirical assessments confirm that the FACD outperforms contemporary denoising techniques by augmenting the peak signal-to-noise ratio by approximately 2.5 dB, underscoring its superior robustness. Furthermore, in light of its specificity to this noise model, FACD rapidly denoises high-resolution real infrared linear array scans, thus meeting the stringent real-time and resolution imperatives of advanced infrared linear array scanning apparatuses.

Real-Time Recognition Method for Key Signals of Rock Fracture Acoustic Emissions Based on Deep Learning

The characteristics of acoustic emission signals generated in the process of rock deformation and fission contain rich information on internal rock damage. The use of acoustic emissions monitoring technology can analyze and identify the precursor information of rock failure. At present, in the field of acoustic emissions monitoring and the early warning of rock fracture disasters, there is no real-time identification method for a disaster precursor characteristic signal. It is easy to lose information by analyzing the characteristic parameters of traditional acoustic emissions to find signals that serve as precursors to disasters, and analysis has mostly been based on post-analysis, which leads to poor real-time recognition of disaster precursor characteristics and low application levels in the engineering field. Based on this, this paper regards the acoustic emissions signal of rock fracture as a kind of speech signal generated by rock fracture uses this idea of speech recognition for reference alongside spectral analysis (STFT) and Mel frequency analysis to realize the feature extraction of acoustic emissions from rock fracture. In deep learning, based on the VGG16 convolutional neural network and AlexNet convolutional neural network, six intelligent real-time recognition models of rock fracture and key acoustic emission signals were constructed, and the network structure and loss function of traditional VGG16 were optimized. The experimental results show that these six deep-learning models can achieve the real-time intelligent recognition of key signals, and Mel, combined with the improved VGG16, achieved the best performance with 87.68% accuracy and 81.05% recall. Then, by comparing multiple groups of signal recognition models, Mel+VGG-FL proposed in this paper was verified as having a high recognition accuracy and certain recognition efficiency, performing the intelligent real-time recognition of key acoustic emission signals in the process of rock fracture more accurately, which can provide new ideas and methods for related research and the real-time intelligent recognition of rock fracture precursor characteristics.

Quantifying innervation facilitated by deep learning in wound healing

The peripheral nerves (PNs) innervate the dermis and epidermis, and are suggested to play an important role in wound healing. Several methods to quantify skin innervation during wound healing have been reported. Those usually require multiple observers, are complex and labor-intensive, and the noise/background associated with the immunohistochemistry (IHC) images could cause quantification errors/user bias. In this study, we employed the state-of-the-art deep neural network, Denoising Convolutional Neural Network (DnCNN), to perform pre-processing and effectively reduce the noise in the IHC images. Additionally, we utilized an automated image analysis tool, assisted by Matlab, to accurately determine the extent of skin innervation during various stages of wound healing. The 8 mm wound is generated using a circular biopsy punch in the wild-type mouse. Skin samples were collected on days 3, 7, 10 and 15, and sections from paraffin-embedded tissues were stained against pan-neuronal marker- protein-gene-product 9.5 (PGP 9.5) antibody. On day 3 and day 7, negligible nerve fibers were present throughout the wound with few only on the lateral boundaries of the wound. On day 10, a slight increase in nerve fiber density appeared, which significantly increased on day 15. Importantly, we found a positive correlation (R2 = 0.926) between nerve fiber density and re-epithelization, suggesting an association between re-innervation and re-epithelization. These results established a quantitative time course of re-innervation in wound healing, and the automated image analysis method offers a novel and useful tool to facilitate the quantification of innervation in the skin and other tissues.

A Simple Denoising Algorithm for Real-World Noisy Camera Images

The noise statistics of real-world camera images are challenging for any denoising algorithm. Here, I describe a modified version of a bionic algorithm that improves the quality of real-word noisy camera images from a publicly available image dataset. In the first step, an adaptive local averaging filter was executed for each pixel to remove moderate sensor noise while preserving fine image details and object contours. In the second step, image sharpness was enhanced by means of an unsharp mask filter to generate output images that are close to ground-truth images (multiple averages of static camera images). The performance of this denoising algorithm was compared with five popular denoising methods: bm3d, wavelet, non-local means (NL-means), total variation (TV) denoising and bilateral filter. Results show that the two-step filter had a performance that was similar to NL-means and TV filtering. Bm3d had the best denoising performance but sometimes led to blurry images. This novel two-step filter only depends on a single parameter that can be obtained from global image statistics. To reduce computation time, denoising was restricted to the Y channel of YUV-transformed images and four image segments were simultaneously processed in parallel on a multi-core processor.

Shot noise reduction in radiographic and tomographic multi-channel imaging with self-supervised deep learning

Shot noise is a critical issue in radiographic and tomographic imaging, especially when additional constraints lead to a significant reduction of the signal-to-noise ratio. This paper presents a method for improving the quality of noisy multi-channel imaging datasets, such as data from time or energy-resolved imaging, by exploiting structural similarities between channels. To achieve that, we broaden the application domain of the Noise2Noise self-supervised denoising approach. The method draws pairs of samples from a data distribution with identical signals but uncorrelated noise. It is applicable to multi-channel datasets if adjacent channels provide images with similar enough information but independent noise. We demonstrate the applicability and performance of the method via three case studies, namely spectroscopic X-ray tomography, energy-dispersive neutron tomography, and in vivo X-ray cine-radiography.

Transfer Learning Approach to Vascular Permeability Changes in Brain Metastasis Post-Whole-Brain Radiotherapy

The purpose of this study is to further validate the utility of our previously developed CNN in an alternative small animal model of BM through transfer learning. Unlike the glioma model, the BM mouse model develops multifocal intracranial metastases, including both contrast enhancing and non-enhancing lesions on DCE MRI, thus serving as an excellent brain tumor model to study tumor vascular permeability. Here, we conducted transfer learning by transferring the previously trained GBM CNN to DCE MRI datasets of BM mice. The CNN was re-trained to learn about the relationship between BM DCE images and target permeability maps extracted from the Extended Tofts Model (ETM). The transferred network was found to accurately predict BM permeability and presented with excellent spatial correlation with the target ETM PK maps. The CNN model was further tested in another cohort of BM mice treated with WBRT to assess vascular permeability changes induced via radiotherapy. The CNN detected significantly increased permeability parameter Ktrans in WBRT-treated tumors (p < 0.01), which was in good agreement with the target ETM PK maps. In conclusion, the proposed CNN can serve as an efficient and accurate tool for characterizing vascular permeability and treatment responses in small animal brain tumor models.

Deconvoluting Kinetic Rate Constants of Catalytic Substrates from Scanning Electrochemical Approach Curves with Artificial Neural Networks

Extracting information from experimental measurements in the chemical sciences typically requires curve fitting, deconvolution, and/or solving the governing partial differential equations via numerical (e.g., finite element analysis) or analytical methods. However, using numerical or analytical methods for high-throughput data analysis typically requires significant postprocessing efforts. Here, we show that deep learning artificial neural networks can be a very effective tool for extracting information from experimental data. As an example, reactivity and topography information from scanning electrochemical microscopy (SECM) approach curves are highly convoluted. This study utilized multilayer perceptrons and convolutional neural networks trained on simulated SECM data to extract kinetic rate constants of catalytic substrates. Our key findings were that multilayer perceptron models performed very well when the experimental data were close to the ideal conditions with which the model was trained. However, convolutional neural networks, which analyze images as opposed to direct data, were able to accurately predict the kinetic rate constant of Fe-doped nickel (oxy)hydroxide catalyst at different applied potentials even though the experimental approach curves were not ideal. Due to the speed at which machine learning models can analyze data, we believe this study shows that artificial neural networks could become powerful tools in high-throughput data analysis.

Simultaneous Super-Resolution and Classification of Lung Disease Scans

Acute lower respiratory infection is a leading cause of death in developing countries. Hence, progress has been made for early detection and treatment. There is still a need for improved diagnostic and therapeutic strategies, particularly in resource-limited settings. Chest X-ray and computed tomography (CT) have the potential to serve as effective screening tools for lower respiratory infections, but the use of artificial intelligence (AI) in these areas is limited. To address this gap, we present a computer-aided diagnostic system for chest X-ray and CT images of several common pulmonary diseases, including COVID-19, viral pneumonia, bacterial pneumonia, tuberculosis, lung opacity, and various types of carcinoma. The proposed system depends on super-resolution (SR) techniques to enhance image details. Deep learning (DL) techniques are used for both SR reconstruction and classification, with the InceptionResNetv2 model used as a feature extractor in conjunction with a multi-class support vector machine (MCSVM) classifier. In this paper, we compare the proposed model performance to those of other classification models, such as Resnet101 and Inceptionv3, and evaluate the effectiveness of using both softmax and MCSVM classifiers. The proposed system was tested on three publicly available datasets of CT and X-ray images and it achieved a classification accuracy of 98.028% using a combination of SR and InceptionResNetv2. Overall, our system has the potential to serve as a valuable screening tool for lower respiratory disorders and assist clinicians in interpreting chest X-ray and CT images. In resource-limited settings, it can also provide a valuable diagnostic support.

Gamma Radiation Image Noise Prediction Method Based on Statistical Analysis and Random Walk

The gamma radiation environment is one of the harshest operating environments for image acquisition systems, and the captured images are heavily noisy. In this paper, we improve the multi-frame difference method for the characteristics of noise and add an edge detection algorithm to segment the noise region and extract the noise quantization information. A Gaussian mixture model of the gamma radiation noise is then established by performing a specific statistical analysis of the amplitude and quantity information of the noise. The established model is combined with the random walk algorithm to generate noise and achieve the prediction of image noise under different accumulated doses. Evaluated by objective similarity matching, there is no significant difference between the predicted image noise and the actual noise in subjective perception. The ratio of similarity-matched images in the sample from the predicted noise to the actual noise reaches 0.908. To further illustrate the spillover effect of this research, in the discussion session, we used the predicted image noise as the training set input to a deep residual network for denoising. The network model was able to achieve a good denoising effect. The results show that the prediction method proposed in this paper can accomplish the prediction of gamma radiation image noise, which is beneficial to the elimination of image noise in this environment.

Object tracking in infrared images using a deep learning model and a target-attention mechanism

Small object tracking in infrared images is widely utilized in various fields, such as video surveillance, infrared guidance, and unmanned aerial vehicle monitoring. The existing small target detection strategies in infrared images suffer from submerging the target in heavy cluttered infrared (IR) maritime images. To overcome this issue, we use the original image and the corresponding encoded image to apply our model. We use the local directional number patterns algorithm to encode the original image to represent more unique details. Our model is able to learn more informative and unique features from the original and encoded image for visual tracking. In this study, we explore the best convolutional filters to obtain the best possible visual tracking results by finding those inactive to the backgrounds while active in the target region. To this end, the attention mechanism for the feature extracting framework is investigated comprising a scale-sensitive feature generation component and a discriminative feature generation module based on the gradients of regression and scoring losses. Comprehensive experiments have demonstrated that our pipeline obtains competitive results compared to recently published papers.

Dual Autoencoder Network with Separable Convolutional Layers for Denoising and Deblurring Images

A dual autoencoder employing separable convolutional layers for image denoising and deblurring is represented. Combining two autoencoders is presented to gain higher accuracy and simultaneously reduce the complexity of neural network parameters by using separable convolutional layers. In the proposed structure of the dual autoencoder, the first autoencoder aims to denoise the image, while the second one aims to enhance the quality of the denoised image. The research includes Gaussian noise (Gaussian blur), Poisson noise, speckle noise, and random impulse noise. The advantages of the proposed neural network are the number reduction in the trainable parameters and the increase in the similarity between the denoised or deblurred image and the original one. The similarity is increased by decreasing the main square error and increasing the structural similarity index. The advantages of a dual autoencoder network with separable convolutional layers are demonstrated by a comparison of the proposed network with a convolutional autoencoder and dual convolutional autoencoder.

Validation of a Denoising Method Using Deep Learning-Based Reconstruction to Quantify Multiple Sclerosis Lesion Load on Fast FLAIR Imaging

BACKGROUND AND PURPOSE

Accurate quantification of WM lesion load is essential for the care of patients with multiple sclerosis. We tested whether the combination of accelerated 3D-FLAIR and denoising using deep learning-based reconstruction could provide a relevant strategy while shortening the imaging examination.

MATERIALS AND METHODS

Twenty-eight patients with multiple sclerosis were prospectively examined using 4 implementations of 3D-FLAIR with decreasing scan times (4 minutes 54 seconds, 2 minutes 35 seconds, 1 minute 40 seconds, and 1 minute 15 seconds). Each FLAIR sequence was reconstructed without and with denoising using deep learning-based reconstruction, resulting in 8 FLAIR sequences per patient. Image quality was assessed with the Likert scale, apparent SNR, and contrast-to-noise ratio. Manual and automatic lesion segmentations, performed randomly and blindly, were quantitatively evaluated against ground truth using the absolute volume difference, true-positive rate, positive predictive value, Dice similarity coefficient, Hausdorff distance, and F1 score based on the lesion count. The Wilcoxon signed-rank test and 2-way ANOVA were performed.

RESULTS

Both image-quality evaluation and the various metrics showed deterioration when the FLAIR scan time was accelerated. However, denoising using deep learning-based reconstruction significantly improved subjective image quality and quantitative performance metrics, particularly for manual segmentation. Overall, denoising using deep learning-based reconstruction helped to recover contours closer to those from the criterion standard and to capture individual lesions otherwise overlooked. The Dice similarity coefficient was equivalent between the 2-minutes-35-seconds-long FLAIR with denoising using deep learning-based reconstruction and the 4-minutes-54-seconds-long reference FLAIR sequence.

CONCLUSIONS

Denoising using deep learning-based reconstruction helps to recognize multiple sclerosis lesions buried in the noise of accelerated FLAIR acquisitions, a possibly useful strategy to efficiently shorten the scan time in clinical practice.

Diagnosis of liver disease by computer- assisted imaging techniques: A literature review

Diagnosis of liver disease using computer-aided detection (CAD) systems is one of the most efficient and cost-effective methods of medical image diagnosis. Accurate disease detection by using ultrasound images or other medical imaging modalities depends on the physician’s or doctor’s experience and skill. CAD systems have a critical role in helping experts make accurate and right-sized assessments. There are different types of CAD systems for diagnosing different diseases, and one of the applications is in liver disease diagnosis and detection by using intelligent algorithms to detect any abnormalities. Machine learning and deep learning algorithms and models play also a big role in this area. In this article, we tried to review the techniques which are utilized in different stages of CAD systems and pursue the methods used in preprocessing, extracting, and selecting features and classification. Also, different techniques are used to segment and analyze the liver ultrasound medical images, which is still a challenging approach to how to use these techniques and their technical and clinical effectiveness as a global approach.

An Infrared Array Sensor-Based Approach for Activity Detection, Combining Low-Cost Technology with Advanced Deep Learning Techniques

In this paper, we propose an activity detection system using a 24 × 32 resolution infrared array sensor placed on the ceiling. We first collect the data at different resolutions (i.e., 24 × 32, 12 × 16, and 6 × 8) and apply the advanced deep learning (DL) techniques of Super-Resolution (SR) and denoising to enhance the quality of the images. We then classify the images/sequences of images depending on the activities the subject is performing using a hybrid deep learning model combining a Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM). We use data augmentation to improve the training of the neural networks by incorporating a wider variety of samples. The process of data augmentation is performed by a Conditional Generative Adversarial Network (CGAN). By enhancing the images using SR, removing the noise, and adding more training samples via data augmentation, our target is to improve the classification accuracy of the neural network. Through experiments, we show that employing these deep learning techniques to low-resolution noisy infrared images leads to a noticeable improvement in performance. The classification accuracy improved from 78.32% to 84.43% (for images with 6 × 8 resolution), and from 90.11% to 94.54% (for images with 12 × 16 resolution) when we used the CNN and CNN + LSTM networks, respectively.

Nonlinear Noise Cleaning in Gravitational-Wave Detectors With Convolutional Neural Networks

Currently, the sub-60 Hz sensitivity of gravitational-wave (GW) detectors like Advanced LIGO (aLIGO) is limited by the control noises from auxiliary degrees of freedom which nonlinearly couple to the main GW readout. One promising way to tackle this challenge is to perform nonlinear noise mitigation using convolutional neural networks (CNNs), which we examine in detail in this study. In many cases, the noise coupling is bilinear and can be viewed as a few fast channels' outputs modulated by some slow channels. We show that we can utilize this knowledge of the physical system and adopt an explicit “slow×fast” structure in the design of the CNN to enhance its performance of noise subtraction. We then examine the requirements in the signal-to-noise ratio (SNR) in both the target channel (i.e., the main GW readout) and in the auxiliary sensors in order to reduce the noise by at least a factor of a few. In the case of limited SNR in the target channel, we further demonstrate that the CNN can still reach a good performance if we use curriculum learning techniques, which in reality can be achieved by combining data from quiet times and those from periods with active noise injections.

Reconstructing high fidelity digital rock images using deep convolutional neural networks

Imaging methods have broad applications in geosciences. Scanning electron microscopy (SEM) and micro-CT scanning have been applied for studying various geological problems. Despite significant advances in imaging capabilities, and image processing algorithms, acquiring high-quality data from images is still challenging and time-consuming. Obtaining a 3D representative volume for a tight rock sample takes days to weeks. Image artifacts such as noise further complicate the use of imaging methods for the determination of rock properties. In this study, we present applications of several convolutional neural networks (CNN) for rapid image denoising, deblurring and super-resolving digital rock images. Such an approach enables rapid imaging of larger samples, which in turn improves the statistical relevance of the subsequent analysis. We demonstrate the application of several CNNs for image restoration applicable to scientific imaging. The results show that images can be denoised without a priori knowledge of the noise with great confidence. Furthermore, we show how attaching several CNNs in an end-to-end fashion can improve the final quality of reconstruction. Our experiments with SEM and CT scan images of several rock types show image denoising, deblurring and super-resolution can be performed simultaneously.

A Convolutional Autoencoder Topology for Classification in High-Dimensional Noisy Image Datasets

Deep convolutional neural networks have shown remarkable performance in the image classification domain. However, Deep Learning models are vulnerable to noise and redundant information encapsulated into the high-dimensional raw input images, leading to unstable and unreliable predictions. Autoencoders constitute an unsupervised dimensionality reduction technique, proven to filter out noise and redundant information and create robust and stable feature representations. In this work, in order to resolve the problem of DL models' vulnerability, we propose a convolutional autoencoder topological model for compressing and filtering out noise and redundant information from initial high dimensionality input images and then feeding this compressed output into convolutional neural networks. Our results reveal the efficiency of the proposed approach, leading to a significant performance improvement compared to Deep Learning models trained with the initial raw images.

QQ-NET - using deep learning to solve quantitative susceptibility mapping and quantitative blood oxygen level dependent magnitude (QSM+qBOLD or QQ) based oxygen extraction fraction (OEF) mapping

PURPOSE

To improve accuracy and speed of quantitative susceptibility mapping plus quantitative blood oxygen level-dependent magnitude (QSM+qBOLD or QQ) -based oxygen extraction fraction (OEF) mapping using a deep neural network (QQ-NET).

METHODS

The 3D multi-echo gradient echo images were acquired in 34 ischemic stroke patients and 4 healthy subjects. Arterial spin labeling and diffusion weighted imaging (DWI) were also performed in the patients. NET was developed to solve the QQ model inversion problem based on Unet. QQ-based OEF maps were reconstructed with previously introduced temporal clustering, tissue composition, and total variation (CCTV) and NET. The results were compared in simulation, ischemic stroke patients, and healthy subjects using a two-sample Kolmogorov-Smirnov test.

RESULTS

In the simulation, QQ-NET provided more accurate and precise OEF maps than QQ-CCTV with 150 times faster reconstruction speed. In the subacute stroke patients, OEF from QQ-NET had greater contrast-to-noise ratio (CNR) between DWI-defined lesions and their unaffected contralateral normal tissue than with QQ-CCTV: 1.9 ± 1.3 vs 6.6 ± 10.7 (p = 0.03). In healthy subjects, both QQ-CCTV and QQ-NET provided uniform OEF maps.

CONCLUSION

QQ-NET improves the accuracy of QQ-based OEF with faster reconstruction.