Active deep learning from a noisy teacher for semi-supervised 3D image segmentation: Application to COVID-19 pneumonia infection in CT.

Supervised deep learning has become a standard approach to solving medical image segmentation tasks. However, serious difficulties in attaining pixel-level annotations for sufficiently large volumetric datasets in real-life applications have highlighted the critical need for alternative approaches, such as semi-supervised learning, where model training can leverage small expert-annotated datasets to enable learning from much larger datasets without laborious annotation. Most of the semi-supervised approaches combine expert annotations and machine-generated annotations with equal weights within deep model training, despite the latter annotations being relatively unreliable and likely to affect model optimization negatively. To overcome this, we propose an active learning approach that uses an example re-weighting strategy, where machine-annotated samples are weighted (i) based on the similarity of their gradient directions of descent to those of expert-annotated data, and (ii) based on the gradient magnitude of the last layer of the deep model. Specifically, we present an active learning strategy with a query function that enables the selection of reliable and more informative samples from machine-annotated batch data generated by a noisy teacher. When validated on clinical COVID-19 CT benchmark data, our method improved the performance of pneumonia infection segmentation compared to the state of the art.

Learning-to-augment incorporated noise-robust deep CNN for detection of COVID-19 in noisy X-ray images.

Deep convolutional neural networks (CNNs) are used for the detection of COVID-19 in X-ray images. The detection performance of deep CNNs may be reduced by noisy X-ray images. To improve the robustness of a deep CNN against impulse noise, we propose a novel CNN approach using adaptive convolution, with the aim to ameliorate COVID-19 detection in noisy X-ray images without requiring any preprocessing for noise removal. This approach includes an impulse noise-map layer, an adaptive resizing layer, and an adaptive convolution layer to the conventional CNN framework. We also used a learning-to-augment strategy using noisy X-ray images to improve the generalization of a deep CNN. We have collected a dataset of 2093 chest X-ray images including COVID-19 (452 images), non-COVID pneumonia (621 images), and healthy ones (1020 images). The architecture of pre-trained networks such as SqueezeNet, GoogleNet, MobileNetv2, ResNet18, ResNet50, ShuffleNet, and EfficientNetb0 has been modified to increase their robustness to impulse noise. Validation on the noisy X-ray images using the proposed noise-robust layers and learning-to-augment strategy-incorporated ResNet50 showed 2% better classification accuracy compared with state-of-the-art method.

Greedy Autoaugment for classification of mycobacterium tuberculosis image via generalized deep CNN using mixed pooling based on minimum square rough entropy.

Although tuberculosis (TB) is a disease whose cause, epidemiology and treatment are well known, some infected patients in many parts of the world are still not diagnosed by current methods, leading to further transmission in society. Creating an accurate image-based processing system for screening patients can help in the early diagnosis of this disease. We provided a dataset containing1078 confirmed negative and 469 positive Mycobacterium tuberculosis instances. An effective method using an improved and generalized convolutional neural network (CNN) was proposed for classifying TB bacteria in microscopic images. In the preprocessing phase, the insignificant parts of microscopic images are excluded with an efficient algorithm based on the square rough entropy (SRE) thresholding. Top 10 policies of data augmentation were selected with the proposed model based on the Greedy AutoAugment algorithm to resolve the overfitting problem. In order to improve the generalization of CNN, mixed pooling was used instead of baseline one. The results showed that employing generalized pooling, batch normalization, Dropout, and PReLU have improved the classification of Mycobacterium tuberculosis images. The output of classifiers such as Naïve Bayes-LBP, KNN-LBP, GBT-LBP, Naïve Bayes-HOG, KNN-HOG, SVM-HOG, GBT-HOG indicated that proposed CNN has the best results with an accuracy of 93.4%. The improvements of CNN based on the proposed model can yield promising results for diagnosing TB.

Learning-to-augment strategy using noisy and denoised data: Improving generalizability of deep CNN for the detection of COVID-19 in X-ray images.

Chest X-ray images are used in deep convolutional neural networks for the detection of COVID-19, the greatest human challenge of the 21st century. Robustness to noise and improvement of generalization are the major challenges in designing these networks. In this paper, we introduce a strategy for data augmentation using the determination of the type and value of noise density to improve the robustness and generalization of deep CNNs for COVID-19 detection. Firstly, we present a learning-to-augment approach that generates new noisy variants of the original image data with optimized noise density. We apply a Bayesian optimization technique to control and choose the optimal noise type and its parameters. Secondly, we propose a novel data augmentation strategy, based on denoised X-ray images, that uses the distance between denoised and original pixels to generate new data. We develop an autoencoder model to create new data using denoised images corrupted by the Gaussian and impulse noise. A database of chest X-ray images, containing COVID-19 positive, healthy, and non-COVID pneumonia cases, is used to fine-tune the pre-trained networks (AlexNet, ShuffleNet, ResNet18, and GoogleNet). The proposed method performs better results compared to the state-of-the-art learning to augment strategies in terms of sensitivity (0.808), specificity (0.915), and F-Measure (0.737). The source code of the proposed method is available at https://github.com/mohamadmomeny/Learning-to-augment-strategy.

Ecological and health risk assessment of exposure to atmospheric heavy metals.

In the present study, we assessed the concentration of airborne HMs (Zn, Cu, Pb, and Cd) and their probable sources using the bark of Pinus eldarica as a bio-indicator. Hence, 47 tree bark samples were harvested according to the land uses and biomonitoring techniques in the city of Yazd, Iran. The potential health risks in 13 age groups, ecological risk, as well as the possible relationship between HM concentrations and traffic indicators, were evaluated. The order of average HM concentrations in the P. eldarica bark samples was as Zn > Pb > Cu > Cd. The mean values of non-carcinogenic risks of all HMs in entire age groups were within secure range (HQ < 1); however, the carcinogenic risk of Cd was higher than the allowed level (TCR > 1 × 10-6). About Pb, it was in the safe level. The main element causing potential ecological risks was Cd, indicating moderate to very high ecological risk in most of the study areas. There was an inverse significant association between distance from major roads and Pb concentration (ß = -0.011 95% confidence interval (CI): 0.022, -0.0001). All HMs in bark samples render the negative Moran's index, representing a random spatial distribution pattern. Besides, according to principal component analysis (PCA), the first component accounted for 36.55% of the total variance, dominated by Cd, Pb, Cu, and Zn, respectively, and characterized by vehicle and industrial emissions. Our results infer that industrial activities and traffic are the main sources of HMs pollution in urban environments that should be considered by decision-makers.

Seroepidemiology of Toxocara canis in Children under 14 Years Referring to Laboratories of Sistan and Baluchestan Province in Southeast of Iran.

BACKGROUND: The aim of the present survey was to assess thr seroepidemiologic and parasitological aspects of Toxocara canis infection in children under 14 yr old. METHODS: Overall, 963 sera were collected from children in the Sistan and Baluchistan Province, Southeast of Iran during the period from Sep 2015 to Jun 2016. IgG antibody against T. canis in the subjects' sera was evaluated using the commercial ELISA kit. RESULTS: Anti-Toxocara IgG were detected in the serum of 17 (1.7%) of the participants. In the examined children, the highest presence of anti-Toxocara antibodies was 2.1% (9/418) in 6-10-yr olds, which was higher than other age groups (P<0.05). Anti-Toxocara antibodies were significantly higher in males (2.4% or 12/492) than in females (1.1% or 5/471) (P<0.03). Highest serological prevalence of T. canis occurred in tribes (5.5% or 4/69), followed by rural areas (0.9% or 7/757), while in the urban area it was 0.1% (6/163) (P<0.01). A significant association was seen between the serological prevalence of T. canis and laboratory findings such as eosinophilia (P=0.001) and red blood cell count (P=0.02). CONCLUSION: Seroprevalence of Toxocara infection is high among children living in the poor regions of southeast Iran.

The role of VEGF in melanoma progression.

BACKGROUND: Melanoma is the most serious skin cancer. There is an established correlation between thickness and aggressiveness of the tumor. Nevertheless, the potential value of vascular endothelial growth factor (VEGF) in correlation with tumor progression remains unresolved. MATERIALS AND METHODS: Thirty seven paraffin blocks of cutaneous melanoma were obtained from Pathology department of Al-zahra hospital between 2005 and 2010. The sections were stained with monoclonal mouse antibodies (mAbs) against vascular endothelial growth factor A and evaluated by distribution of expression of VEGF in tumor cells as 0, 0%; 1, 1%--25%; 2, 25%--50%; 3, >50% and the staining intensity from 0 (negative) to 3 (strong). The sum of intensity score and distribution score was then calculated as the VEGF index. The relationship between VEGF expression (distribution, intensity, and index) and tumor progression (vertical and radial growth, Clark's level, and Breslow's depth) was studied. SPSS software was used to analyze the data by ANOVA, and chi-square tests. RESULTS: 51.4% of the patients showed vertical growth pattern. Mean Breslow's depth was 1.84 ± 1.79 mm. There was a significant association between growth pattern and VEGF distribution, intensity and index (P = 0.006, P = 0.005, and P = 0.001 respectively). VEGF distribution, intensity, and index all had correlation with Breslow's depth as well (ANOVA test: P = 0.003, P < 0.001, and P < 0.001 respectively) VEGF index had also correlation with Clark's level, but this was not seen for VEGF distribution and intensity. CONCLUSION: VEGF expression (both VEGF distribution and intensity) is associated with progression of malignant melanoma. VEGF index can explain this association better.