MPCNN: A Novel Matrix Profile Approach for CNN-based Single Lead Sleep Apnea in Classification Problem.

Sleep apnea (SA) is a significant respiratory condition that poses a major global health challenge. Deep Learning (DL) has emerged as an efficient tool for the classification problem in electrocardiogram (ECG)-based SA diagnoses. Despite these advancements, most common conventional feature extractions derived from ECG signals in DL, such as R-peaks and RR intervals, may fail to capture crucial information encompassed within the complete ECG segments. In this study, we propose an innovative approach to address this diagnostic gap by delving deeper into the comprehensive segments of the ECG signal. The proposed methodology draws inspiration from Matrix Profile algorithms, which generate an Euclidean distance profile from fixed-length signal subsequences. From this, we derived the Min Distance Profile (MinDP), Max Distance Profile (MaxDP), and Mean Distance Profile (MeanDP) based on the minimum, maximum, and mean of the profile distances, respectively. To validate the effectiveness of our approach, we use the modified LeNet-5 architecture as the primary CNN model, along with two existing lightweight models, BAFNet and SE-MSCNN. Our experiment results on the PhysioNet Apnea-ECG dataset (70 overnight recordings), and the UCDDB dataset (25 overnight recordings) revealed that our new feature extraction method achieved per-segment accuracies of up to 92.11% and 81.25%, respectively. Moreover, using the PhysioNet data, we achieved a per-recording accuracy of 100% and yielded the highest correlation of 0.989 compared to state-of-the-art methods. By introducing a new feature extraction method based on distance relationships, we enhanced the performance of certain lightweight models in DL, showing potential for home sleep apnea test (HSAT) and SA detection in IoT devices. The source code for this work is made publicly available in GitHub: https://github.com/vinuni-vishc/MPCNN-Sleep-Apnea.

VinDr-Mammo: A large-scale benchmark dataset for computer-aided diagnosis in full-field digital mammography.

Mammography, or breast X-ray imaging, is the most widely used imaging modality to detect cancer and other breast diseases. Recent studies have shown that deep learning-based computer-assisted detection and diagnosis (CADe/x) tools have been developed to support physicians and improve the accuracy of interpreting mammography. A number of large-scale mammography datasets from different populations with various associated annotations and clinical data have been introduced to study the potential of learning-based methods in the field of breast radiology. With the aim to develop more robust and more interpretable support systems in breast imaging, we introduce VinDr-Mammo, a Vietnamese dataset of digital mammography with breast-level assessment and extensive lesion-level annotations, enhancing the diversity of the publicly available mammography data. The dataset consists of 5,000 mammography exams, each of which has four standard views and is double read with disagreement (if any) being resolved by arbitration. The purpose of this dataset is to assess Breast Imaging Reporting and Data System (BI-RADS) and breast density at the individual breast level. In addition, the dataset also provides the category, location, and BI-RADS assessment of non-benign findings. We make VinDr-Mammo publicly available as a new imaging resource to promote advances in developing CADe/x tools for mammography interpretation.

PediCXR: An open, large-scale chest radiograph dataset for interpretation of common thoracic diseases in children.

Computer-aided diagnosis systems in adult chest radiography (CXR) have recently achieved great success thanks to the availability of large-scale, annotated datasets and the advent of high-performance supervised learning algorithms. However, the development of diagnostic models for detecting and diagnosing pediatric diseases in CXR scans is undertaken due to the lack of high-quality physician-annotated datasets. To overcome this challenge, we introduce and release PediCXR, a new pediatric CXR dataset of 9,125 studies retrospectively collected from a major pediatric hospital in Vietnam between 2020 and 2021. Each scan was manually annotated by a pediatric radiologist with more than ten years of experience. The dataset was labeled for the presence of 36 critical findings and 15 diseases. In particular, each abnormal finding was identified via a rectangle bounding box on the image. To the best of our knowledge, this is the first and largest pediatric CXR dataset containing lesion-level annotations and image-level labels for the detection of multiple findings and diseases. For algorithm development, the dataset was divided into a training set of 7,728 and a test set of 1,397. To encourage new advances in pediatric CXR interpretation using data-driven approaches, we provide a detailed description of the PediCXR data sample and make the dataset publicly available on https://physionet.org/content/vindr-pcxr/1.0.0/ .

Detecting COVID-19 from digitized ECG printouts using 1D convolutional neural networks.

The COVID-19 pandemic has exposed the vulnerability of healthcare services worldwide, raising the need to develop novel tools to provide rapid and cost-effective screening and diagnosis. Clinical reports indicated that COVID-19 infection may cause cardiac injury, and electrocardiograms (ECG) may serve as a diagnostic biomarker for COVID-19. This study aims to utilize ECG signals to detect COVID-19 automatically. We propose a novel method to extract ECG signals from ECG paper records, which are then fed into one-dimensional convolution neural network (1D-CNN) to learn and diagnose the disease. To evaluate the quality of digitized signals, R peaks in the paper-based ECG images are labeled. Afterward, RR intervals calculated from each image are compared to RR intervals of the corresponding digitized signal. Experiments on the COVID-19 ECG images dataset demonstrate that the proposed digitization method is able to capture correctly the original signals, with a mean absolute error of 28.11 ms. The 1D-CNN model (SEResNet18), which is trained on the digitized ECG signals, allows to identify between individuals with COVID-19 and other subjects accurately, with classification accuracies of 98.42% and 98.50% for classifying COVID-19 vs. Normal and COVID-19 vs. other classes, respectively. Furthermore, the proposed method also achieves a high-level of performance for the multi-classification task. Our findings indicate that a deep learning system trained on digitized ECG signals can serve as a potential tool for diagnosing COVID-19.

Learning to diagnose common thorax diseases on chest radiographs from radiology reports in Vietnamese.

Deep learning, in recent times, has made remarkable strides when it comes to impressive performance for many tasks, including medical image processing. One of the contributing factors to these advancements is the emergence of large medical image datasets. However, it is exceedingly expensive and time-consuming to construct a large and trustworthy medical dataset; hence, there has been multiple research leveraging medical reports to automatically extract labels for data. The majority of this labor, however, is performed in English. In this work, we propose a data collecting and annotation pipeline that extracts information from Vietnamese radiology reports to provide accurate labels for chest X-ray (CXR) images. This can benefit Vietnamese radiologists and clinicians by annotating data that closely match their endemic diagnosis categories which may vary from country to country. To assess the efficacy of the proposed labeling technique, we built a CXR dataset containing 9,752 studies and evaluated our pipeline using a subset of this dataset. With an F1-score of at least 0.9923, the evaluation demonstrates that our labeling tool performs precisely and consistently across all classes. After building the dataset, we train deep learning models that leverage knowledge transferred from large public CXR datasets. We employ a variety of loss functions to overcome the curse of imbalanced multi-label datasets and conduct experiments with various model architectures to select the one that delivers the best performance. Our best model (CheXpert-pretrained EfficientNet-B2) yields an F1-score of 0.6989 (95% CI 0.6740, 0.7240), AUC of 0.7912, sensitivity of 0.7064 and specificity of 0.8760 for the abnormal diagnosis in general. Finally, we demonstrate that our coarse classification (based on five specific locations of abnormalities) yields comparable results to fine classification (twelve pathologies) on the benchmark CheXpert dataset for general anomaly detection while delivering better performance in terms of the average performance of all classes.

A novel multi-view deep learning approach for BI-RADS and density assessment of mammograms.

Advanced deep learning (DL) algorithms may predict the patient's risk of developing breast cancer based on the Breast Imaging Reporting and Data System (BI-RADS) and density standards. Recent studies have suggested that the combination of multi-view analysis improved the overall breast exam classification. In this paper, we propose a novel multi-view DL approach for BI-RADS and density assessment of mammograms. The proposed approach first deploys deep convolutional networks for feature extraction on each view separately. The extracted features are then stacked and fed into a Light Gradient Boosting Machine (LightGBM) classifier to predict BI-RADS and density scores. We conduct extensive experiments on both the internal mammography dataset and the public dataset Digital Database for Screening Mammogra-phy (DDSM). The experimental results demonstrate that the proposed approach outperforms the single-view classification approach on two benchmark datasets by huge F1-score margins (+5% on the internal dataset and +10% on the DDSM dataset). These results highlight the vital role of combining multi-view information to improve the performance of breast cancer risk prediction.

VinDr-CXR: An open dataset of chest X-rays with radiologist's annotations.

Most of the existing chest X-ray datasets include labels from a list of findings without specifying their locations on the radiographs. This limits the development of machine learning algorithms for the detection and localization of chest abnormalities. In this work, we describe a dataset of more than 100,000 chest X-ray scans that were retrospectively collected from two major hospitals in Vietnam. Out of this raw data, we release 18,000 images that were manually annotated by a total of 17 experienced radiologists with 22 local labels of rectangles surrounding abnormalities and 6 global labels of suspected diseases. The released dataset is divided into a training set of 15,000 and a test set of 3,000. Each scan in the training set was independently labeled by 3 radiologists, while each scan in the test set was labeled by the consensus of 5 radiologists. We designed and built a labeling platform for DICOM images to facilitate these annotation procedures. All images are made publicly available in DICOM format along with the labels of both the training set and the test set.

Phase recognition in contrast-enhanced CT scans based on deep learning and random sampling.

PURPOSE: A fully automated system for interpreting abdominal computed tomography (CT) scans with multiple phases of contrast enhancement requires an accurate classification of the phases. Current approaches to classify the CT phases are commonly based on three-dimensional (3D) convolutional neural network (CNN) approaches with high computational complexity and high latency. This work aims at developing and validating a precise, fast multiphase classifier to recognize three main types of contrast phases in abdominal CT scans. METHODS: We propose in this study a novel method that uses a random sampling mechanism on top of deep CNNs for the phase recognition of abdominal CT scans of four different phases: noncontrast, arterial, venous, and others. The CNNs work as a slicewise phase prediction, while random sampling selects input slices for the CNN models. Afterward, majority voting synthesizes the slicewise results of the CNNs to provide the final prediction at the scan level. RESULTS: Our classifier was trained on 271 426 slices from 830 phase-annotated CT scans, and when combined with majority voting on 30% of slices randomly chosen from each scan, achieved a mean F1 score of 92.09% on our internal test set of 358 scans. The proposed method was also evaluated on two external test sets: CTPAC-CCRCC (N = 242) and LiTS (N = 131), which were annotated by our experts. Although a drop in performance was observed, the model performance remained at a high level of accuracy with a mean F1 scores of 76.79% and 86.94% on CTPAC-CCRCC and LiTS datasets, respectively. Our experimental results also showed that the proposed method significantly outperformed the state-of-the-art 3D approaches while requiring less computation time for inference. CONCLUSIONS: In comparison to state-of-the-art classification methods, the proposed approach shows better accuracy with significantly reduced latency. Our study demonstrates the potential of a precise, fast multiphase classifier based on a two-dimensional deep learning approach combined with a random sampling method for contrast phase recognition, providing a valuable tool for extracting multiphase abdomen studies from low veracity, real-world data.

The novel method to reduce the silica content in lignin recovered from black liquor originating from rice straw.

Difficulties in the production of lignin from rice straw because of high silica content in the recovered lignin reduce its recovery yield and applications as bio-fuel and aromatic chemicals. Therefore, the objective of this study is to develop a novel method to reduce the silica content in lignin from rice straw more effectively and selectively. The method is established by monitoring the precipitation behavior as well as the chemical structure of precipitate by single-stage acidification at different pH values of black liquor collected from the alkaline treatment of rice straw. The result illustrates the significant influence of pH on the physical and chemical properties of the precipitate and the supernatant. The simple two-step acidification of the black liquor at pilot-scale by sulfuric acid 20w/v% is applied to recover lignin at pH 9 and pH 3 and gives a percentage of silica removal as high as 94.38%. Following the developed process, the high-quality lignin could be produced from abundant rice straw at the industrial-scale.

Effects of the c-Si/a-SiO2 interfacial atomic structure on its band alignment: an ab initio study.

The crystalline-Si/amorphous-SiO2 (c-Si/a-SiO2) interface is an important system used in many applications, ranging from transistors to solar cells. The transition region of the c-Si/a-SiO2 interface plays a critical role in determining the band alignment between the two regions. However, the question of how this interface band offset is affected by the transition region thickness and its local atomic arrangement is yet to be fully investigated. Here, by controlling the parameters of the classical Monte Carlo bond switching algorithm, we have generated the atomic structures of the interfaces with various thicknesses, as well as containing Si at different oxidation states. A hybrid functional method, as shown by our calculations to reproduce the GW and experimental results for bulk Si and SiO2, was used to calculate the electronic structure of the heterojunction. This allowed us to study the correlation between the interface band characterization and its atomic structures. We found that although the systems with different thicknesses showed quite different atomic structures near the transition region, the calculated band offset tended to be the same, unaffected by the details of the interfacial structure. Our band offset calculation agrees well with the experimental measurements. This robustness of the interfacial electronic structure to its interfacial atomic details could be another reason for the success of the c-Si/a-SiO2 interface in Si-based electronic applications. Nevertheless, when a reactive force field is used to generate the a-SiO2 and c-Si/a-SiO2 interfaces, the band offset significantly deviates from the experimental values by about 1 eV.

DFT+U studies of Cu doping and p-type compensation in crystalline and amorphous ZnS.

Zinc sulfide is an excellent candidate for the development of a p-type transparent conducting material that has great demands in solar energy and optoelectronic applications. Doping with Cu is one potential way to make ZnS p-type while preserving its optical transparency for the solar spectrum; however, this is limited by the extremely low solubility of Cu in ZnS and charge compensation mechanisms that eliminate the p-type characteristics. These mechanisms are different in crystalline (c-ZnS) and amorphous structures (a-ZnS), leading to different tendencies of doping Cu in these two ZnS phases, as well as the feasibility to form the p-type material. In this work, we have carried out fundamental studies of Cu doping in both c-ZnS and a-ZnS, using the continuous random network model and density functional theory with Hubbard's energy correction (DFT+U). The formation of a complex that contains two CuZn and one S vacancy is highly favorable in both phases. The local environment of this charge-compensated Cu complex obtained by DFT calculations agrees well with the previous EXAFS measurements. The incorporation of Cu into a-ZnS, on the one hand, is more tolerable compared to its crystal counterparts (zincblende), indicating possible higher Cu concentration. On the other hand, there is also another intrinsic mechanism to compensate the p-type characteristics in a-ZnS: the formation of the covalent S-S "dumbbell" units. This reconstruction of the local structure to form a S-S bond could occur spontaneously, thus making the p-type doping for ZnS challenging even in the amorphous phase.

Electronic structures and current conductivities of B, C, N and F defects in amorphous titanium dioxide.

Although titanium dioxide (TiO2) has been extensively studied and widely used in energy and environmental areas, the amorphous form and its related defect properties are poorly understood. Recent studies, however, have emphasized the crucial role of amorphousness in producing competitively good performances in photochemical applications. In this work we have investigated for the first time the effects of various dopants (B, C, N and F) on charge carrier transport in amorphous titanium dioxide (a-TiO2), given that doping is a common technique used to tune the electronic properties of semiconductors, and that the existence of these impurities could also be unintentionally introduced during the synthesis process. The a-TiO2 model was obtained using a classical molecular dynamics method, followed by density-functional theory calculations (DFT + U, with Hubbard correction term U) on electronic structures and defect states. The formation of these impurity defects in a-TiO2 was found to be energetically more favorable by several eV than their crystal counterparts (in rutile). The contributions of these defect states to the charge transfer processes were examined by means of Marcus theory.

Oxygen vacancy and hole conduction in amorphous TiO2.

The amorphous titanium dioxide (a-TiO2) has drawn attention recently due to the finding that it holds promise for coating conventional photoelectrodes for corrosion protection while still allowing the holes to transport to the surface. The mechanism of hole conductivity at a level much higher than the edge of the valence band is still a mystery. In this work, an amorphous TiO2 model is obtained from molecular dynamics employing the "melt-and-quench" technique. The electronic properties, polaronic states and the hole conduction mechanism in amorphous structure were investigated by means of density functional theory with Hubbard's energy correction (DFT + U) and compared to those in crystalline (rutile) TiO2. The formation energy of the oxygen vacancy was found to reduce significantly (by a few eV) upon amorphization. Our theoretical study suggested that the oxygen vacancies and their defect states provide hopping channels, which are comparable to experimental observations and could be responsible for hole conduction in the "leaky" TiO2 recently discovered for the photochemical water-splitting applications.

First-principles prediction of the effects of temperature and solvent selection on the dimerization of benzoic acid.

We introduce a procedure of quantum chemical calculations (B3P86/6-31G**) to study carboxylic acid dimerization and its correlation with temperature and properties of the solvent. Benzoic acid is chosen as a model system for studying dimerization via hydrogen bonding. Organic solvents are simulated using the self-consistent reaction field (SCRF) method with the polarized continuum model (PCM). The cyclic dimer is the most stable structure both in gas phase and solution. Dimer mono- and dihydrates could be found in the gas phase if acid molecules are in contact with water vapor. However, the formation of these hydrated conformers is very limited and cyclic dimer is the principal conformer to coexist with monomer acid in solution. Solvation of the cyclic dimer is more favorable compared to other complexes, partially due to the diminishing of hydrogen bonding capability and annihilation of dipole moments. Solvents have a strong effect on inducing dimer dissociation and this dependence is more pronounced at low dielectric constants. By accounting for selected terms in the total free energy of solvation, the solvation entropy could be incorporated to predict the dimer behavior at elevated temperatures. The temperature dependence of benzoic acid dimerization obtained by this technique is in good agreement with available experimental measurements, in which a tendency of dimer to dissociate is observed with increased temperatures. In addition, dimer breakup is more sensitive to temperature in low dielectric environments rather than in solvents with a higher dielectric constant.