Search

Abstract This study presents a novel privacy-preserving self-supervised (SSL) framework for COVID-19 classification from lung CT scans, utilizing federated learning (FL) enhanced with Paillier homomorphic encryption (PHE) to prevent third-party attacks during training. The FL-SSL based framework employs two publicly available lung CT scan datasets which are considered as labeled and an unlabeled dataset. The unlabeled dataset is split into three subsets which are assumed to be collected from three hospitals. Training is done using the Bootstrap Your Own Latent (BYOL) contrastive learning SSL framework with a VGG19 encoder followed by attention CNN blocks (VGG19 + attention CNN). The input datasets are processed by selecting the largest lung portion of each lung CT scan using an automated selection approach and a 64 × 64 input size is utilized to reduce computational complexity. Healthcare privacy issues are addressed by collaborative training across decentralized datasets and secure aggregation with PHE, underscoring the effectiveness of this approach. Three subsets of the dataset are used to train the local BYOL model, which together optimizes the central encoder. The labeled dataset is employed to train the central encoder (updated VGG19 + attention CNN), resulting in an accuracy of 97.19%, a precision of 97.43%, and a recall of 98.18%. The reliability of the framework’s performance is demonstrated through statistical analysis and five-fold cross-validation. The efficacy of the proposed framework is further showcased by showing its performance on three distinct modality datasets: skin cancer, breast cancer, and chest X-rays. In conclusion, this study offers a promising solution for accurate diagnosis of chest X-rays, preserving privacy and overcoming the challenges of dataset scarcity and computational complexity.

Endometrial cancеr is the fourth fastеst-growing cancеr among women worldwide, affecting the uterus's lining. This research proposes a novel approach called ECgMLP for the automated diagnosis of endometrial cancer by analyzing histopathological images. Several preprocessing techniques are employed to increase the quality of the images, including normalization, Non-Local Means denoising, and alpha-beta enhancement. Effective segmentation is achieved through a combination of Otsu thresholding, morphological operations, distance transformations, and the watershed approach to identify major regions of interest. Through a sequence of blocks, the ECgMLP architecture processes input images to remove unimportant patterns. Model hyperparameters are improved via ablation research. The evaluations show a maximum accuracy of 99.26 % for identifying multi-class histopathological categories of endometrial tissue, which is higher than the previous best technique. The proposed model offers an automated, correct diagnosis, enhancing clinical processes. This proposition could be added to the current tools for finding endometrial cancer early, leading to better patient outcomes.

A novel automated multi-classification approach is proposed for the anticipation of lung abnormalities using chest X-ray and CT images. The study leverages a publicly accessible dataset with an insufficient and unbalanced number of images, addressing this issue by employing the data augmentation approach DCGAN to balance the dataset. Various preprocessing procedures are applied to improve features and reduce noise in lung pictures. As the base for the model, the vision trans-former and convolution-based compact convolutional transformer (CCT) model is utilized. To determine the best model configuration, an ablation study is performed on the original CCT model using a CT scan dataset with image dimensions of $32\times32$ . Following that, this model is trained on the X-ray dataset to evaluate performance on an entirely other modality. The performances are compared to six pre-trained models with $32\times 32$ images. While traditional models achieved modest performance, with test accuracies ranging from 43% to 77% and 49% to 73% requiring lengthy training times, the suggested model performed exceptionally well, obtaining test accuracies of 99.77% and 95.37% for CT and X-ray, respectively with a short training duration of 10–12 and 40–42 seconds/epoch. Robustness is demonstrated through the progressive reduction of the number of training images, with findings indicating that the model maintains good performance even on a reduced dataset. An explainable AI technique Grad-CAM is used to explain the model’s judgment. Grad-CAM-based color visualization is shown to explain model assessments and help health specialists make quick, confident decisions. This study used image preprocessing and deep learning techniques to detect lung anomalies, and it addressed the challenges of training time and computational complexity.

Autism spectrum disorder (ASD) is a complex neurodevelopmental condition that presents significant diagnostic challenges due to its varied symptoms and nature. This study aims to improve ASD classification using advanced deep learning techniques applied to neuroimaging data. We developed an automated system leveraging the ABIDE-1 dataset and a novel lightweight quantized one-dimensional (1D) Convolutional Neural Network (Q-CNN) model to analyze fMRI data. Our approach employs the NIAK pipeline with multiple brain atlases and filtering methods. Initially, the Regions of Interest (ROIs) are converted into feature vectors using tangent space embedding to feed into the Q-CNN model. The proposed 1D-CNN is quantized through Quantize Aware Training (QAT). As the quantization method, int8 quantization is utilized, which makes it both robust and lightweight. We propose a federated learning (FL) framework to ensure data privacy, which allows decentralized training across different data centers without compromising local data security. Our findings indicate that the CC200 brain atlas, within the NIAK pipeline’s filt-global filtering methods, provides the best results for ASD classification. Notably, the ASD classification outcomes have achieved a significant test accuracy of 98% using the CC200 and filt-global filtering techniques. To the best of our knowledge, this performance surpasses previous studies in the field, highlighting a notable enhancement in ASD detection from fMRI data. Furthermore, the FL-based Q-CNN model demonstrated robust performance and high efficiency on a Raspberry Pi 4, underscoring its potential for real-world applications. We exhibit the efficacy of the Q-CNN model by comparing its inference time, power consumption, and storage requirements with those of the 1D-CNN, quantized CNN, and the proposed int8 Q-CNN models. This research has made several key contributions, including the development of a lightweight int8 Q-CNN model, the application of FL for data privacy, and the evaluation of the proposed model in real-world settings. By identifying optimal brain atlases and filtering methods, this study provides valuable insights for future research in the field of neurodevelopmental disorders.

Objective Early diagnosis of breast cancer can lead to effective treatment, possibly increase long-term survival rates, and improve quality of life. The objective of this study is to present an automated analysis and classification system for breast cancer using clinical markers such as tumor shape, orientation, margin, and surrounding tissue. The novelty and uniqueness of the study lie in the approach of considering medical features based on the diagnosis of radiologists. Methods Using clinical markers, a graph is generated where each feature is represented by a node, and the connection between them is represented by an edge which is derived through Pearson's correlation method. A graph convolutional network (GCN) model is proposed to classify breast tumors into benign and malignant, using the graph data. Several statistical tests are performed to assess the importance of the proposed features. The performance of the proposed GCN model is improved by experimenting with different layer configurations and hyper-parameter settings. Results Results show that the proposed model has a 98.73% test accuracy. The performance of the model is compared with a graph attention network, a one-dimensional convolutional neural network, and five transfer learning models, ten machine learning models, and three ensemble learning models. The performance of the model was further assessed with three supplementary breast cancer ultrasound image datasets, where the accuracies are 91.03%, 94.37%, and 89.62% for Dataset A, Dataset B, and Dataset C (combining Dataset A and Dataset B) respectively. Overfitting issues are assessed through k-fold cross-validation. Conclusion Several variants are utilized to present a more rigorous and fair evaluation of our work, especially the importance of extracting clinically relevant features. Moreover, a GCN model using graph data can be a promising solution for an automated feature-based breast image classification system.

Bangladesh's Haor regions are famous for their natural resources and are unable to escape climate vulnerability. Triggered by climate vulnerabilities farmers are heading towards climate-resilient approaches. Hence, research was done in the haor area of Sunamganj district to analyze the status of adopted Climate-Smart Agriculture (CSA) techniques in Chhatak, Sunamganj, and Jagannathpur which are prone to severe flooding and climate conditions. Around 450 farmers were randomly selected and CSA adopters were contacted. A structured questionnaire was prepared with open-ended and closed-ended questions. The final questionnaire contained demographic questions and a list of adopted cropland and homestead CSA practices, and the survey proceeded with 115 finalized CSA adopters. MS Excel and SPSS were used to analyze the data. The data were expressed using frequency, percent, mean, and standard deviation. A t-test, analysis of variance, multiple linear regression, Pearson correlation, boxplot, and normal P–P plots were employed to test data normality. The analysis revealed that 30 CSA practices were identified to be practiced in cropland where major preferences were found for appropriate seed storage (100%), USG application (100%), IPM (98%), and good quality seed (95%) in cropland, whereas agroforestry (71%), organic fertilizer application (63%), perching (63%) and IPM (59%) were major CSA practices among the 18 identified practices in homesteads. The adoption level of CSA practices was found in the score category of 11–23 for cropland (90%) and up to 10 for homestead (68%). The results showed that the adoption status of CSA practices was inefficient for quick flood occurrence. CSA practices are not applied enough in haor areas' homesteads due to lack of knowledge, information access, and technical and financial resources. Thus, CSA should be implemented which necessitates working on barriers restricting CSA adoption through strengthening the infrastructure of technologies, supportive policies, and institutional framework.

Searching, retrieving, and arranging text in ever-larger document collections necessitate more efficient information processing algorithms. Document categorization is a crucial component of various information processing systems for supervised learning. As the quantity of documents grows, the performance of classic supervised classifiers has deteriorated because of the number of document categories. Assigning documents to a predetermined set of classes is called text classification. It is utilized extensively in a wide range of data-intensive applications. However, the fact that real-world implementations of these models are plagued with shortcomings begs for more investigation. Imbalanced datasets hinder the most prevalent high-performance algorithms. In this paper, we propose an approach name multi-class Convolutional Neural Network (MCNN)-Long Short-Time Memory (LSTM), which combines two deep learning techniques, Convolutional Neural Network (CNN) and Long Short-Time Memory, for text classification in news data. CNN’s are used as feature extractors for the LSTMs on text input data and have the spatial structure of words in a sentence, paragraph, or document. The dataset is also imbalanced, and we use the Tomek-Link algorithm to balance the dataset and then apply our model, which shows better performance in terms of F1-score (98%) and Accuracy (99.71%) than the existing works. The combination of deep learning techniques used in our approach is ideal for the classification of imbalanced datasets with underrepresented categories. Hence, our method outperformed other machine learning algorithms in text classification by a large margin. We also compare our results with traditional machine learning algorithms in terms of imbalanced and balanced datasets.

Work-related musculoskeletal ailments are injuries or disorders of the joints, muscles, nerves, or tendons caused by repetitive tasks and jobs that require uncomfortable postures. REBA (Rapid Entire Body Assessment) is a widely used assessment method for examining occupational ergonomics in areas where musculoskeletal disorders (MSDs) are common. REBA assessment necessitates the presence of a professional evaluator who monitors workers’ motions and postures, which takes time and has limitations in terms of real-world implementation. With the progress of deep learning-based human posture estimate algorithms, postural risk assessment has become an important and complex research area. We present a technique for forecasting REBA risk levels using 3D coordinates of human body position as input data in this study. We calculated REBA risk scores for various body segments and overall risk rating for corresponding action level for each body position using 3D keypoints from the widely renowned Human 3.6M dataset, which is a significant contribution for future research work in this arena. Using this vast ground truth dataset, a unique DNN model was created to forecast the REBA risk level for measuring the full body’s postural risk. REBA Ground Truth dataset is highly imbalanced which coped with data augmentation for the rare classes. To determine the optimal model configuration based on highest accuracy, ablation study is conducted by tuning different hyper-parameters. The proposed model, post-ablation study, attained 89.07% accuracy score on a test set of 128,046 samples from Nadam optimizer with a learning rate of 0.001 and batch size of 512.

Bronchiectasis in children is a major health issue which can be life-threatening if not diagnosed and effectively treated. In the diagnosis of bronchiectasis, an increased broncho-arterial (BA) ratio is considered a significant marker. The BA ratio is measured by evaluating BA pairs, using high-resolution computed tomography (HRCT) scans. Detecting BA pairs automatically is challenging due to the complex characteristics of BA pairs and the ambiguous appearance of the bronchi. This study proposes an effective computerized approach to detect BA pairs and assess BA ratio using HRCT scans of children and employing computer-aided techniques and novel custom-build algorithms. Attention is given to reconstructing broken bronchial walls and identifying discrete BA pairs using custom-built kernel based and patch-based algorithms for pixel-level image analysis. To detect BA pairs, the lung region is segmented in the HRCT slices and image preprocessing techniques, including noise reduction, binarizing, largest contour detection and a hole-filling algorithm, are applied. A histogram analysis method is introduced to clean the images. A kernel-based algorithm is proposed to reconstruct the pixel distribution if the bronchial wall is so that the bronchi can be detected precisely. Potential arteries are detected using balanced histogram thresholding, morphological opening and an approach based on four conditions related to the object area circularity, rectangular boundary box ratio and enclosing circle area ratio. Potential bronchi are detected through matching of object coordinates with potential arteries, hole-filling and four condition based approaches. The potential BA pairs are detected by matching the coordinates of potential bronchi with those of potential arteries as the artery and bronchus are adjacent to each other in BA pairs. Finally, from the potential BA pairs, actual BA pairs are identified using a custom-built patch algorithm. The study is conducted using 2471 HRCT slices of seven children, obtained from the Royal Darwin Hospital, Australia. The BA ratio is derived based on the ratio of diameters, major axis lengths, minor axis lengths, area, convex hull and equivalent diameter where the BA ratios are respectively 0.51–0.65, 0.49–0.59, 0.59–0.77, 0.25–0.42, 0.29–0.47, 1.5–2 and 0.50–0.65.

COVID-19, pneumonia, and tuberculosis have had a significant effect on recent global health. Since 2019, COVID-19 has been a major factor underlying the increase in respiratory-related terminal illness. Early-stage interpretation and identification of these diseases from X-ray images is essential to aid medical specialists in diagnosis. In this study, (COV-X-net19) a convolutional neural network model is developed and customized with a soft attention mechanism to classify lung diseases into four classes: normal, COVID-19, pneumonia, and tuberculosis using chest X-ray images. Image preprocessing is carried out by adjusting optimal parameters to preprocess the images before undertaking training of the classification models. Moreover, the proposed model is optimized by experimenting with different architectural structures and hyperparameters to further boost performance. The performance of the proposed model is compared with eight state-of-the-art transfer learning models for a comparative evaluation. Results suggest that the COV-X-net19 outperforms other models with a testing accuracy of 95.19%, precision of 96.49% and F1-score of 95.13%. Another novel approach of this study is to find out the probable reason behind image misclassification by analyzing the handcrafted imaging features with statistical evaluation. A statistical analysis known as analysis of variance test is performed, to identify at which point the model can identify a class accurately, and at which point the model cannot identify the class. The potential features responsible for the misclassification are also found. Moreover, Random Forest Feature importance technique and Minimum Redundancy Maximum Relevance technique are also explored. The methods and findings of this study can benefit in the clinical perspective in early detection and enable a better understanding of the cause of misclassification.

Introduction: Breast cancer stands as the second most deadly form of cancer among women worldwide. Early diagnosis and treatment can significantly mitigate mortality rates. Purpose: The study aims to classify breast ultrasound images into benign and malignant tumors. This approach involves segmenting the breast's region of interest (ROI) employing an optimized UNet architecture and classifying the ROIs through an optimized shallow CNN model utilizing an ablation study. Method: Several image processing techniques are utilized to improve image quality by removing text, artifacts, and speckle noise, and statistical analysis is done to check the enhanced image quality is satisfactory. With the processed dataset, the segmentation of breast tumor ROI is carried out, optimizing the UNet model through an ablation study where the architectural configuration and hyperparameters are altered. After obtaining the tumor ROIs from the fine-tuned UNet model (RKO-UNet), an optimized CNN model is employed to classify the tumor into benign and malignant classes. To enhance the CNN model's performance, an ablation study is conducted, coupled with the integration of an attention unit. The model's performance is further assessed by classifying breast cancer with mammogram images. Result: The proposed classification model (RKONet-13) results in an accuracy of 98.41 %. The performance of the proposed model is further compared with five transfer learning models for both pre-segmented and post-segmented datasets. K-fold cross-validation is done to assess the proposed RKONet-13 model's performance stability. Furthermore, the performance of the proposed model is compared with previous literature, where the proposed model outperforms existing methods, demonstrating its effectiveness in breast cancer diagnosis. Lastly, the model demonstrates its robustness for breast cancer classification, delivering an exceptional performance of 96.21 % on a mammogram dataset. Conclusion: The efficacy of this study relies on image pre-processing, segmentation with hybrid attention UNet, and classification with fine-tuned robust CNN model. This comprehensive approach aims to determine an effective technique for detecting breast cancer within ultrasound images.

Growing energy demand has exacerbated the issue of energy security and caused us to necessitate the utilization of renewable resources. The best alternative for promoting generation in Bangladesh from renewable energy is solar photovoltaic technology. Grid-connected solar photovoltaic (PV) systems are becoming increasingly popular, considering solar potential and the recent cost of PV modules. This study proposes a grid-connected solar PV system with a net metering strategy using the Hybrid Optimization of Multiple Electric Renewables model. The HOMER model is used to evaluate raw data, to create a demand cycle using data from load surveys, and to find the best cost-effective configuration. A sensitivity analysis was also conducted to assess the impact of differences in radiation from the solar (4, 4.59, 4.65, 5 kWh/m2/day), PV capacity (0 kW, 100 kW, 200 kW, 300 kW, 350 kW, 400 kW, 420 kW), and grid prices (＄0.107, ＄0.118, ＄0.14 per kWh) upon that optimum configuration. Outcomes reveal that combining 420 kW of PV with a 405-kW converter and connecting to the utility grid is the least expensive and ecologically healthy configuration of the system. The electricity generation cost is estimated to be 0.0725 dollars per kilowatt-hour, and the net present value is 1.83 million dollars with a payback period of 6.4 years based on the system's 20-year lifespan. Also, compared to the existing grid and diesel-generator system, the optimized system, with a renewable fraction of 31.10%, provides a reduction in carbon dioxide emissions of 191 tons and 1,028 tons, respectively, each year.

Convolutional neural networks (CNNs) have been established for a comprehensive range of computer vision problems across several benchmarks. Visualization and analysis of feature maps generated by convolutional layers can be an effective approach to explore the hidden and complex characteristic of a CNN model. Convolutional layers provide diverse feature maps however, the extent of this diversity needs to be explored. This research attempts to provide five insights of the ‘Black box’ mechanism of CNNs, using skin cancer dermoscopy and lung scan computed tomography (CT) Scan datasets by statistically analyzing layer by layer (three convolutional layers) feature maps using 17 geometrical and 6 intensity-based features to determine the characteristics and level of diversity. Significance and difference of the feature maps layer by layer, black feature maps analysis, difference of the feature maps to each other and to the original image, variations among the feature maps when running the model multiple times and inter-class variation among the feature maps for different iteration are explored. Various statistical methods including T-test, analysis of variance (ANOVA), mean, median, mean squared error (MSE), peak signal to noise ratio (PSNR), structural similarity index (SSIM), root mean squared error (RMSE), dice similarity score (DSC), universal image quality index (UQI) and Spectral angle mapper (SAM) are employed. Experimental results show that for the skin cancer dermoscopy dataset, a large number of black feature maps are produced (20–60%) while the proportion of black feature maps for the CT Scan dataset is comparatively low (2–20%). This demonstrates that for different datasets, feature maps with diverse characteristics can be produced. The layer by layer differences between the feature maps is evaluated using T-tests and ANOVA for seventeen geometrical features and six intensity-based features. For both datasets across most of the geometrical features and across most of the intensity-based features a significant diversity can be observed. The difference of the feature maps to each other and to the original image is quite high, with MSE values for the dermoscopy and CT Scan datasets in the range of 1860–31,399 and 171–6089, respectively, PSNR 3–15 and 10–25, SSIM values of 0.01–0.84 and 0.3–0.81, RMSE values of 0.81–1 and 0.21–1, DSC values of 0.37–0.53 and 0.47–0.75, UQI values of 0.02–0.86 and 0.01–0.88 and SAM values of 0.12–1.53 and 0.19–1.55 for the dermoscopy and CT Scan datasets respectively. When running the model multiple times (three iterations), a notable iteration by iteration diversity is found in terms of mean, median, maximum and minimum values for most of the geometrical features. The inter-class variation among the feature maps for different iterations and layers are evaluated based on the F-value of the ANOVA test. For the dermoscopy dataset, the highest mean F-value is found for layer 1 and iteration 3 while for the CT scan dataset the highest mean F-value is found for layer 3 and iteration 3 indicating that for these feature maps the highest inter-class dissimilarity is generated. The findings of this study may aid in exploring the complex mechanism of convolutional layers, kernels and feature maps.

Identification of brain tumors at an early stage is crucial in cancer diagnosis, as a timely diagnosis can increase the chances of survival. Considering the challenges of tumor biopsies, three dimensional (3D) Magnetic Resonance Imaging (MRI) are extensively used in analyzing brain tumors using deep learning. In this study, three BraTS datasets are employed to classify brain tumor into two classes where each of the datasets contains four 3D MRI sequences for a single patient. This research is composed of two approaches. In the first part, we propose a hybrid model named TimeDistributed-CNN-LSTM (TD- CNN-LSTM) combining 3D Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) where each layer is wrapped with a TimeDistributed function. The objective is to consider all the four MRI sequences of each patient as a single input data because every sequence contains necessary information of tumor. Therefore, the model is developed with optimal configuration performing ablation study for layer architecture and hyper-parameters. In the second part, a 3D CNN model is trained respectively with each of the MRI sequences to compare the performance. Moreover, the datasets are preprocessed to ensure highest performance. Results demonstrate that the TD-CNN-LSTM network outperforms 3D CNN achieving the highest test accuracy of 98.90%. Later, to evaluate the performance consistency, the TD-CNN-LSTM model is evaluated with K-fold cross validation. The approach of putting together all the MRI sequences at a time with good generalization capability can be used in future medical research which can aid radiologists in tumor diagnostics effectively.

Named Entity Recognition (NER) plays a significant role in enhancing the performance of all types of domain specific applications in Natural Language Processing (NLP). According to the type of application, the goal of NER is to identify target entities based on the context of other existing entities in a sentence. Numerous architectures have demonstrated good performance for high-resource languages such as English and Chinese NER. However, currently existing NER models for Bengali could not achieve reliable accuracy due to morphological richness of Bengali and limited availability of resources. This work integrates both Data and Model Centric AI concepts to achieve a state-of-the-art performance. A unique dataset was created for this study demonstrating the impact of a good quality dataset on accuracy. We proposed a method for developing a high quality NER dataset for any language. We have used our dataset to evaluate the performance of various Deep Learning models. A hybrid model performed with the exact match F1 score of 87.50%, partial match F1 score of 92.31%, and micro F1 score of 98.32%. Our proposed model reduces the need for feature engineering and utilizes minimal resources.

Early identification and adequate treatment can help prevent lung disorders from becoming chronic, severe, and life-threatening. X-ray images are commonly used and an automated and effective method involving deep learning techniques can potentially contribute to quick and accurate diagnosis of lung disorders. However, in the study of medical imaging using deep learning, two obstacles limit interpretability. One is an insufficient and imbalanced number of training samples in most medical datasets. The other is excessive training time. Although training time can be reduced by decreasing the number of pixels in the images, training with low resolution images tends to result in poor performance. This study represents a solution to overcome these impediments by balancing the number of images and reducing overall processing time while preserving accuracy. The dataset used in this research contains an unequal number of images in the different classes. The quantity of data in the classes is balanced by creating synthetic images based on the patterns and characteristics of the original images, using a Deep Convolutional Generative Adversarial Network (DCGAN). Unwanted regions are removed from the X-ray images, the brightness and contrast of the images are enhanced, and the abnormalities are highlighted by using different artifact removal, noise reduction, and enhancement techniques. We propose a Modified Compact Convolutional Transformer (MCCT) model using 32 × 32 sized images for the categorization of lung disorders into four classes. An ablation study of eleven cases is employed to adjust several hyper parameters and layer topologies. This reduces training time while preserving accuracy. Six transfer learning models, VGG19, VGG16, ResNet152, ResNet50, ResNet50V2, and MobileNet are applied with the same image size the performance is compared with the proposed MCCT model. Our MCCT model records the greatest test accuracy of 95.37%, requiring a short training time, 10-12 s/epoch, whereas the other models only reach near-moderate performance with accuracies ranging from 43% to 79% and training times of 80-90 s/epoch. The robustness of the model with regards to the number of training samples is validated by training the model multiple times reducing the number of training images gradually from 49621 images to 6204 images. Results suggest that even with a smaller dataset, the performance is sustained. Our proposed approach may contribute to an effective CAD based diagnostic system by addressing the issues of insufficient and imbalanced numbers of medical images, excessive training times and low-resolution images.

Bronchiectasis in children can progress to a severe lung condition if not diagnosed and treated early. The radiological diagnostic criteria for the diagnosis of bronchiectasis is an increased broncho-arterial (BA) ratio. From high-resolution computed tomography (HRCT) scans, the BA pairs must be detected first to derive the BA ratio. This study aims to identify potential BA pairs from HRCT scans of children undertaken to evaluate suppurative lung disease through an automated approach. After segmenting the lung regions, the HRCT scans are cleaned using a histogram analysis-based approach followed by a potential arteries identification process comprising four conditions based on imaging features. Potential arteries and their connected components are extracted, and potential bronchi are identified. Finally, the coordinates of potential arteries and potential bronchi are matched as the last step of BA pairs extraction. A total of 8–50 BA pairs are detected for each patient. Additionally, the area and several diameters of the bronchi and arteries are measured, and BA ratios based on these are calculated. Through this approach, the BA pairs of a CT scan datasets are detected and utilizing a deep learning model, a high classification test accuracy of 98.53% is achieved, validating the robustness of the proposed BA detection approach. The results show that visible BA pairs can be identified and segmented automatically, and the BA ratio calculated may help diagnose bronchiectasis with less effort and time.

Interpretation of medical images with a computer-aided diagnosis (CAD) system is arduous because of the complex structure of cancerous lesions in different imaging modalities, high degree of resemblance between inter-classes, presence of dissimilar characteristics in intra-classes, scarcity of medical data, and presence of artifacts and noises. In this study, these challenges are addressed by developing a shallow convolutional neural network (CNN) model with optimal configuration performing ablation study by altering layer structure and hyper-parameters and utilizing a suitable augmentation technique. Eight medical datasets with different modalities are investigated where the proposed model, named MNet-10, with low computational complexity is able to yield optimal performance across all datasets. The impact of photometric and geometric augmentation techniques on different datasets is also evaluated. We selected the mammogram dataset to proceed with the ablation study for being one of the most challenging imaging modalities. Before generating the model, the dataset is augmented using the two approaches. A base CNN model is constructed first and applied to both the augmented and non-augmented mammogram datasets where the highest accuracy is obtained with the photometric dataset. Therefore, the architecture and hyper-parameters of the model are determined by performing an ablation study on the base model using the mammogram photometric dataset. Afterward, the robustness of the network and the impact of different augmentation techniques are assessed by training the model with the rest of the seven datasets. We obtain a test accuracy of 97.34% on the mammogram, 98.43% on the skin cancer, 99.54% on the brain tumor magnetic resonance imaging (MRI), 97.29% on the COVID chest X-ray, 96.31% on the tympanic membrane, 99.82% on the chest computed tomography (CT) scan, and 98.75% on the breast cancer ultrasound datasets by photometric augmentation and 96.76% on the breast cancer microscopic biopsy dataset by geometric augmentation. Moreover, some elastic deformation augmentation methods are explored with the proposed model using all the datasets to evaluate their effectiveness. Finally, VGG16, InceptionV3, and ResNet50 were trained on the best-performing augmented datasets, and their performance consistency was compared with that of the MNet-10 model. The findings may aid future researchers in medical data analysis involving ablation studies and augmentation techniques.

The complex feature characteristics and low contrast of cancer lesions, a high degree of inter-class resemblance between malignant and benign lesions, and the presence of various artifacts including hairs make automated melanoma recognition in dermoscopy images quite challenging. To date, various computer-aided solutions have been proposed to identify and classify skin cancer. In this paper, a deep learning model with a shallow architecture is proposed to classify the lesions into benign and malignant. To achieve effective training while limiting overfitting problems due to limited training data, image preprocessing and data augmentation processes are introduced. After this, the 'box blur' down-scaling method is employed, which adds efficiency to our study by reducing the overall training time and space complexity significantly. Our proposed shallow convolutional neural network (SCNN_12) model is trained and evaluated on the Kaggle skin cancer data ISIC archive which was augmented to 16485 images by implementing different augmentation techniques. The model was able to achieve an accuracy of 98.87% with optimizer Adam and a learning rate of 0.001. In this regard, parameter and hyper-parameters of the model are determined by performing ablation studies. To assert no occurrence of overfitting, experiments are carried out exploring k-fold cross-validation and different dataset split ratios. Furthermore, to affirm the robustness the model is evaluated on noisy data to examine the performance when the image quality gets corrupted.This research corroborates that effective training for medical image analysis, addressing training time and space complexity, is possible even with a lightweighted network using a limited amount of training data.

Current research indicates that for the identification of lung disorders, comprising pneumonia and COVID-19, structural distortions of bronchi and arteries (BA) should be taken into account. CT scans are an effective modality to detect lung anomalies. However, anomalies in bronchi and arteries can be difficult to detect. Therefore, in this study, alterations of bronchi and arteries are considered in the classification of lung diseases. Four approaches to highlight these are introduced: (a) a Hessian-based approach, (b) a region-growing algorithm, (c) a clustering-based approach, and (d) a color-coding-based approach. Prior to this, the lungs are segmented, employing several image preprocessing algorithms. The utilized COVID-19 Lung CT scan dataset contains three classes named Non-COVID, COVID, and community-acquired pneumonia, having 6983, 7593, and 2618 samples, respectively. To classify the CT scans into three classes, two deep learning architectures, (a) a convolutional neural network (CNN) and (b) a CNN with long short-term memory (LSTM) and an attention mechanism, are considered. Both these models are trained with the four datasets achieved from the four approaches. Results show that the CNN model achieved test accuracies of 88.52%, 87.14%, 92.36%, and 95.84% for the Hessian, the region-growing, the color-coding, and the clustering-based approaches, respectively. The CNN with LSTM and an attention mechanism model results in an increase in overall accuracy for all approaches with an 89.61%, 88.28%, 94.61%, and 97.12% test accuracy for the Hessian, region-growing, color-coding, and clustering-based approaches, respectively. To assess overfitting, the accuracy and loss curves and k-fold cross-validation technique are employed. The Hessian-based and region-growing algorithm-based approaches produced nearly equivalent outcomes. Our proposed method outperforms state-of-the-art studies, indicating that it may be worthwhile to pay more attention to BA features in lung disease classification based on CT images.

Heart disease can be life-threatening if not detected and treated at an early stage. The electrocardiogram (ECG) plays a vital role in classifying cardiovascular diseases, and often physicians and medical researchers examine paper-based ECG images for cardiac diagnosis. An automated heart disease prediction system might help to classify heart diseases accurately at an early stage. This study aims to classify cardiac diseases into five classes with paper-based ECG images using a deep learning approach with the highest possible accuracy and the lowest possible time complexity. This research consists of two approaches. In the first approach, five deep learning models, InceptionV3, ResNet50, MobileNetV2, VGG19, and DenseNet201, are employed. In the second approach, an integrated deep learning model (InRes-106) is introduced, combining InceptionV3 and ResNet50. This model is developed as a deep convolutional neural network capable of extracting hidden and high-level features from images. An ablation study is conducted on the proposed model altering several components and hyperparameters, improving the performance even further. Before training the model, several image pre-processing techniques are employed to remove artifacts and enhance the image quality. Our proposed hybrid InRes-106 model performed best with a testing accuracy of 98.34%. The InceptionV3 model acquired a testing accuracy of 90.56%, the ResNet50 89.63%, the DenseNet201 88.94%, the VGG19 87.87%, and the MobileNetV2 achieved 80.56% testing accuracy. The model is trained with a k-fold cross-validation technique with different k values to evaluate the robustness further. Although the dataset contains a limited number of complex ECG images, our proposed approach, based on various image pre-processing techniques, model fine-tuning, and ablation studies, can effectively diagnose cardiac diseases.

Background: Breast cancer, behind skin cancer, is the second most frequent malignancy among women, initiated by an unregulated cell division in breast tissues. Although early mammogram screening and treatment result in decreased mortality, differentiating cancer cells from surrounding tissues are often fallible, resulting in fallacious diagnosis. Method: The mammography dataset is used to categorize breast cancer into four classes with low computational complexity, introducing a feature extraction-based approach with machine learning (ML) algorithms. After artefact removal and the preprocessing of the mammograms, the dataset is augmented with seven augmentation techniques. The region of interest (ROI) is extracted by employing several algorithms including a dynamic thresholding method. Sixteen geometrical features are extracted from the ROI while eleven ML algorithms are investigated with these features. Three ensemble models are generated from these ML models employing the stacking method where the first ensemble model is built by stacking ML models with an accuracy of over 90% and the accuracy thresholds for generating the rest of the ensemble models are >95% and >96. Five feature selection methods with fourteen configurations are applied to notch up the performance. Results: The Random Forest Importance algorithm, with a threshold of 0.045, produces 10 features that acquired the highest performance with 98.05% test accuracy by stacking Random Forest and XGB classifier, having a higher than >96% accuracy. Furthermore, with K-fold cross-validation, consistent performance is observed across all K values ranging from 3–30. Moreover, the proposed strategy combining image processing, feature extraction and ML has a proven high accuracy in classifying breast cancer.

Background: Identification and treatment of breast cancer at an early stage can reduce mortality. Currently, mammography is the most widely used effective imaging technique in breast cancer detection. However, an erroneous mammogram based interpretation may result in false diagnosis rate, as distinguishing cancerous masses from adjacent tissue is often complex and error-prone. Methods: Six pre-trained and fine-tuned deep CNN architectures: VGG16, VGG19, MobileNetV2, ResNet50, DenseNet201, and InceptionV3 are evaluated to determine which model yields the best performance. We propose a BreastNet18 model using VGG16 as foundational base, since VGG16 performs with the highest accuracy. An ablation study is performed on BreastNet18, to evaluate its robustness and achieve the highest possible accuracy. Various image processing techniques with suitable parameter values are employed to remove artefacts and increase the image quality. A total dataset of 1442 preprocessed mammograms was augmented using seven augmentation techniques, resulting in a dataset of 11,536 images. To investigate possible overfitting issues, a k-fold cross validation is carried out. The model was then tested on noisy mammograms to evaluate its robustness. Results were compared with previous studies. Results: Proposed BreastNet18 model performed best with a training accuracy of 96.72%, a validating accuracy of 97.91%, and a test accuracy of 98.02%. In contrast to this, VGGNet19 yielded test accuracy of 96.24%, MobileNetV2 77.84%, ResNet50 79.98%, DenseNet201 86.92%, and InceptionV3 76.87%. Conclusions: Our proposed approach based on image processing, transfer learning, fine-tuning, and ablation study has demonstrated a high correct breast cancer classification while dealing with a limited number of complex medical images.