Convolutional Neural Network Classifier Research Articles

How does lead selection impact diagnosis through AI? Implications for continuous monitoring systems using convolutional neural networks

Abstract Background and Objectives Recently, there has been a surge in non-invasive diagnosing methods leveraging the dipolar 12-lead ECG information using Convolutional Neural Networks (CNNs). However, the traditional ECG contains 12 projections of the cardiac dipole in a three-dimensional space can result in redundancy. Furthermore, in continuous monitoring systems such as Holter and wearables, not all 12 leads are consistently available. This study aims to quantify redundancy and determine the best lead configuration for improved classification results using a CNN. It evaluates the level of redundancy in ECG leads and its impact on CNN classification. Methods The degree of redundancy was quantified by means of lead entropies (H) as depicted in Figure 1.A. The standard 12-lead ECG was chosen as the baseline input. From it, various subsets of leads were selected (6, 3, 1 leads) with the objective of attaining maximal orthogonality between leads while minimizing redundancy (Figure 1.B). A 2D-CNN model was trained for multiclass classification: Hypertension (HYP), ST-T interval Changes (STTC), Conduction Disturbances (CD), Myocardial Infarction (MI) and Normal (NORM). The designed inputs were employed to evaluate how redundancy impacts the performance of the CNN (see Figure 1.C). Results Reducing redundancy in input leads led to notable improvements in model prediction for all pathological conditions in terms of Area Under the Curve (AUC). Reducing to 6 and 8 leads were the most suitable input for the CNN (Table 1) increasing by 1.18% in HYP, 0.23% in STTC, 0.52% in CD, and 0.33% in MI. However, a significant reduction in redundancy may result in the loss of essential clinical information for classification, as observed in scenarios involving only 3 or 1 leads. Conclusions When diagnosing pathologies with systems featuring fewer leads, its combinations should aim for maximum orthogonality to reduce redundancy while preserving unique information. By strategically selecting a set of leads with reduced input redundancy, the effectiveness of the standard 12-lead ECG can be improved, thereby enhancing the utility of such continuous monitoring systems for automated diagnosis and also decreasing training and prediction times.

Transfer learning classification of suspicious lesions on breast ultrasound: is there room to avoid biopsies of benign lesions?

BackgroundBreast cancer (BC) is the most common malignancy in women and the second cause of cancer death. In recent years, there has been a strong development in artificial intelligence (AI) applications in medical imaging for several tasks. Our aim was to evaluate the potential of transfer learning with convolutional neural networks (CNNs) in discriminating suspicious breast lesions on ultrasound images.MethodsTransfer learning performances of five different CNNs (Inception V3, Xception, Densenet121, VGG 16, and ResNet50) were evaluated on a public and on an institutional dataset (526 and 392 images, respectively), customizing the top layers for the specific task. Institutional images were contoured by an expert radiologist and processed to feed the CNNs for training and testing. Postimaging biopsies were used as a reference standard for classification. The area under the receiver operating curve (AUROC) was used to assess diagnostic performance.ResultsNetworks performed very well on the public dataset (AUROC 0.938–0.996). The direct generalization to the institutional dataset resulted in lower performances (max AUROC 0.676); however, when tested on BI-RADS 3 and BI-RADS 5 only, results were improved (max AUROC 0.792). Good results were achieved on the institutional dataset (AUROC 0.759–0.818) and, when selecting a threshold of 2% for classification, a sensitivity of 0.983 was obtained for three of five CNNs, with the potential to spare biopsy in 15.3%–18.6% of patients.ConclusionIn conclusion, transfer learning with CNNs may achieve high sensitivity and might be used as a support tool in managing suspicious breast lesions on ultrasound images.Relevance statementTransfer learning is a powerful technique to exploit the performances of well-trained CNNs for image classification. In a clinical scenario, it might be useful for the management of suspicious breast lesions on breast ultrasound, potentially sparing biopsy in a non-negligible number of patients.Key PointsProperly trained CNNs with transfer learning are highly effective in differentiating benign and malignant lesions on breast ultrasound.Setting clinical thresholds increased sensitivity.CNNs might be useful as support tools in managing suspicious lesions on breast ultrasound.Graphical

Quantum classical hybrid convolutional neural networks for breast cancer diagnosis.

The World Health Organization states that early diagnosis is essential to increasing the cure rate for breast cancer, which poses a danger to women's health worldwide. However, the efficacy and cost limitations of conventional diagnostic techniques increase the possibility of misdiagnosis. In this work, we present a quantum hybrid classical convolutional neural network (QCCNN) based breast cancer diagnosis approach with the goal of utilizing quantum computing's high-dimensional data processing power and parallelism to increase diagnosis efficiency and accuracy. When working with large-scale and complicated datasets, classical convolutional neural network (CNN) and other machine learning techniques generally demand a large amount of computational resources and time. Their restricted capacity for generalization makes it challenging to maintain consistent performance across multiple data sets. To address these issues, this paper adds a quantum convolutional layer to the classical convolutional neural network to take advantage of quantum computing to improve learning efficiency and processing speed. Simulation experiments on three breast cancer datasets, GBSG, SEER and WDBC, validate the robustness and generalization of QCCNN and significantly outperform CNN and logistic regression models in classification accuracy. This study not only provides a novel method for breast cancer diagnosis but also achieves a breakthrough in breast cancer diagnosis and promotes the development of medical diagnostic technology.

Development of a city-level surface ozone forecasting system using deep learning techniques and air quality model: Application in eastern China

Utilizing regional air quality models to accurately forecast surface ozone (O3) concentrations, particularly high concentrations, is essential for protecting public health. However, forecasts of air quality model often deviate from site observations due to the limitation of grid resolution and uncertainties from emission sources, meteorological conditions, and chemical reaction mechanisms. Especially, the underestimation is significant under condition of high O3 concentrations. Moreover, such deviations tend to accumulate as forecast lead time increases, compounding the challenges associated with reliable air quality forecast. In this study, we employed AlexNet architecture, a classical convolutional neural network, combined with multiple variables related to meteorology, chemistry, emission and geography to establish a non-linear relationship between grid-scale input variables and site-scale hourly O3 forecast biases in Eastern China, aiming to realize accurate city-level ozone forecast based on a regional air quality prediction model (i.e., Nested Air Quality Prediction Model System, NAQPMS). By assigning weights to high-bias samples and high-concentration samples within the loss function, the proposed Weighted AlexNet model (W_AlexNet) effectively reduced forecast biases and enhanced its capability to predict O3 pollution levels. Compared to NAQPMS, W_AlexNet model demonstrated a 25.71% improvement in RMSE and a 7.17% increase in IOA averagely for hourly O3 (O3-1h) forecasts across four different lead times (24-h, 48-h, 72-h, and 96-h). Notably, W_AlexNet model alleviated the tendency of NAQPMS to underestimate high concentrations and showed a superior performance in improving O3-1h pollution level forecasts, particularly for the 72-h and 96-h lead times. W_AlexNet model can effectively mitigate the bias accumulation effect over increasing lead times, thereby enhancing the reliability of longer-term forecasts. Thus, the W_AlexNet model serves as a post-processing model that can calibrate forecast biases in air quality prediction models, significantly improving the accuracy of O3 high concentration forecasts and providing more precise early warnings of O3 pollution. This underscores its utility in air quality management.

Supervised machine learning of outbred mouse genotypes to predict hepatic immunological tolerance of individuals

It is essential to elucidate the molecular mechanisms underlying liver transplant tolerance and rejection. In cases of mouse liver transplantation between inbred strains, immunological rejection of the allograft is reduced with spontaneous apoptosis without immunosuppressive drugs, which differs from the actual clinical result. This may be because inbred strains are genetically homogeneous and less heterogeneous than others. We exploited outbred CD1 mice, which show highly heterogeneous genotypes among individuals, to search for biomarkers related to immune responses and to construct a model for predicting the outcome of liver allografting. Of the 36 mice examined, 18 died within 3 weeks after transplantation, while the others survived for more than 6 weeks. Whole-exome sequencing of the 36 donors revealed more than 9 million variants relative to the C57BL/6 J reference. We selected 6517 single-nucleotide and indel variants and performed machine learning to determine whether or not we could predict the prognosis of each genotype. Models were built by both deep learning with a one-dimensional convolutional neural network and linear classification and evaluated by leave-one-out cross-validation. Given that one short-lived mouse died early in an accident, the models perfectly predicted the outcome of all individuals, suggesting the importance of genotype collection. In addition, linear classification models provided a list of loci potentially responsible for these responses. The present methods as well as results is likely to be applicable to liver transplantation in humans.

Detection and classification of breast cancer in mammographic images with fine-tuned convolutional neural networks

ABSTRACT Breast cancer is cancer that forms in the cells of the breasts and is a severe health issue that affects many people around the world, especially since it is the most deadly cancer in women. By finding it early and using new treatments, patients can overcome this challenge and get back to a healthier life. This study proposed a procedure to fine-tune the Convolutional Neural Networks (CNN) model with data preprocessing and augmentation in classifying mammogram images called the Hybrid Mammogram Classification and Detection Pipeline (HMCaD). After using CNN for classification because it brings higher confidence in classifying tasks, the YOLOv8 has been applied for localization subtask to detect abnormal positions with predicted bounding boxes. The database is provided by the Mammographic Image Analysis Society (MIAS) and is protected by the United Kingdom. It comprises 330 samples, including 79 benign, 54 malignant, and 207 normal images. As a result, the classification in our model based on the custom EfficientNetB3 model and seam carving technique received a great validation accuracy, test accuracy, and F1 score throughout three scenarios at 96.73%, 97.59%, and 97.58%, respectively. Furthermore, the area under the Receiver Operating Characteristic (ROC) curve also has a surprise result of 0.96 (i.e. AUC = 0.96 ). Moreover, YOLOv8 for detecting abnormal positions in our study achieved 83.22% in Intersection over Union (IoU). This led to the research reaching good results in classifying and detecting breast cancer by considering several performance metrics.

Classification of Rice Grains by Image Processing

General Background: The assessment of the surface quality of pre-treated rice grains is crucial for determining their market acceptability, storage stability, processing quality, and overall consumer satisfaction. Traditional evaluation methods are often time-consuming and yield subjective classifications. Specific Background: The lack of a comprehensive tagged image dataset hinders the application of advanced convolutional neural networks (CNN) for detailed damage classification of healthy rice grains. Knowledge Gap: Existing datasets and methods limit the effective exploration of sophisticated CNN models for categorizing rice types, particularly in identifying subtle damage characteristics. Aims: The study aims to create a robust rice grain classification system using image processing techniques, primarily deep learning algorithms, to improve the classification of rice varieties. Results: Utilizing a dataset of 75,000 images across five widely cultivated rice varieties in Turkey, we achieved classification accuracies of 100% for CNN, 99.95% for deep neural networks (DNN), and 99.87% for artificial neural networks (ANN). Novelty: The proposed approach represents a significant advancement in rice classification technology, employing a combination of image acquisition, feature extraction, and machine learning to streamline the process, effectively addressing the challenges faced in traditional methods. Implications: The findings underscore the potential for improved sorting and grading efficiency in the rice industry, facilitating better market outcomes and consumer satisfaction through enhanced quality control Highlights: Innovative: Uses CNN for accurate rice variety classification. Data-Driven: Analyzes 75,000 images for enhanced quality evaluation. Impactful: Increases efficiency in sorting and grading processes. Keywords: Rice Classification, Image Processing, CNN, Machine Learning, Agricultural Technology

Improving acoustic species identification using data augmentation within a deep learning framework

Convolutional neural networks (CNNs) are effective tools for acoustic classification tasks such as species identification. Large datasets of labelled recordings are required to develop CNN classifiers which can be difficult to obtain, particularly if species are rare or vocalise infrequently. Additionally, data often requires manual labelling which can be time consuming requiring expert analysis. Artificially generating data using augmentation can address these challenges, however the impact of data augmentation on CNN performance is poorly understood and often omitted in bioacoustic studies. Here, we empirically test the impact of CNN architecture and 20 data augmentation methods on classifier performance. We use acoustic identification of 18 small mammal species as a case study of a species group that can be effectively surveyed by acoustic monitoring, but recordings for training data are scarce and difficult to collect. Networks that achieved the highest accuracy across all sample sizes was a 10-layer CNN (96.43 %) and a pre-trained ResNet50 model (96.37 %). Overall, all augmentation effects improved ResNet50 model performance and 17 effects improved Conv10 performance, increasing relative change in accuracy (RCA) by 0.021–0.641. Three augmentation effects negatively impacted Conv10 RCA by −0.042 to −0.182. We also show that adding augmented data when the number of original samples is low has the greatest positive impact on accuracy and this effect was larger with ResNet50 models. Our work demonstrates that using data augmentation where few original samples are available can considerably improve model performance and highlights the potential of augmentation in developing acoustic classifiers for species where data are limited or difficult to obtain.

Advanced classification of hot subdwarf binaries using artificial intelligence techniques and Gaia DR3 data

Hot subdwarf stars are compact blue evolved objects, burning helium in their cores surrounded by a tiny hydrogen envelope. In the Hertzsprung-Russell Diagram they are located by the blue end of the Horizontal Branch. Most models agree on a quite probable common envelope binary evolution scenario in the Red Giant phase. However, the current binarity rate for these objects is yet unsolved, but key, question in this field. This study aims to develop a novel classification method for identifying hot subdwarf binaries within large datasets using Artificial Intelligence techniques and data from the third Gaia data release (GDR3). The results will be compared with those obtained previously using Virtual Observatory techniques on coincident samples. The methods used for hot subdwarf binary classification include supervised and unsupervised machine learning techniques. Specifically, we have used Support Vector Machines (SVM) to classify 3084 hot subdwarf stars based on their colour-magnitude properties. Among these, 2815 objects have Gaia DR3 BP/RP spectra, which were classified using Self-Organizing Maps (SOM) and Convolutional Neural Networks (CNN). In order to ensure spectral quality, previously to SOM and CNN classification, our 2815 BP/RP set were pre-analysed with two different approaches: the cosine similarity technique and the Uniform Manifold Approximation and Projection (UMAP) technique. Additional analysis onto a golden sample of 88 well-defined objects, is also presented. The findings demonstrate a high agreement level (sim 70-90<!PCT!>) with the classifications from the Virtual Observatory Sed Analyzer (VOSA) tool. This shows that the SVM, SOM, and CNN methods effectively classify sources with an accuracy comparable to human inspection or non-AI techniques. Notably, SVM in a radial basis function achieves 70.97<!PCT!> reproducibility for binary targets using photometry, and CNN reaches 84.94<!PCT!> for binary detection using spectroscopy. We also found that the single--binary differences are especially observable on the infrared flux in our Gaia DR3 BP/BR spectra, at wavelengths larger than $ sim $700 nm. We find that all the methods used are in fairly good agreement and are particularly effective to discern between single and binary systems. The agreement is also consistent with the results previously obtained with VOSA. In global terms, considering all quality metrics, CNN is the method that provides the best accuracy. The methods also appear effective for detecting peculiarities in the spectra. While promising, challenges in dealing with uncertain compositions highlight the need for caution, suggesting further research is needed to refine techniques and enhance automated classification reliability, particularly for large-scale surveys.

DETECTION OF ORAL SQUAMOUS CELL CARCINOMA USING PRE-TRAINED DEEP LEARNING MODELS.

Oral squamous cell carcinoma (OSCC), the 13th most common type of cancer, claimed 364,339 lives in 2020. Researchers have established a strong correlation between early detection and better prognosis for this type of cancer. Tissue biopsy, the most common diagnostic method used by doctors, is both expensive and time-consuming. The recent growth in using transfer learning methodologies to aid in medical diagnosis, along with the improved 5-year survival rate from early diagnosis serve as motivation for this study. The aim of the study was to evaluate an innovative approach using transfer learning of pre-trained classification models and convolutional neural networks (CNN) for the binary classification of OSCC from histopathological images. The dataset used for the experiments consisted of 5192 histopathological images in total. The following pre-trained deep learning models were used for feature extraction: ResNet-50, VGG16, and InceptionV3 along with a tuned CNN for classification. The proposed methodologies were evaluated against the current state of the art. A high sensitivity and its importance in the medical field were highlighted. All three models were used in experiments with different hyperparameters and tested on a set of 126 histopathological images. The highest-performance developed model achieved an accuracy of 0.90, a sensitivity of 0.97, and an AUC of 0.94. The visualization of the results was done using ROC curves and confusion matrices. The study further interprets the results obtained and concludes with suggestions for future research. The study successfully demonstrated the potential of using transfer learning-based methodologies in the medical field. The interpretation of the results suggests their practical viability and offers directions for future research aimed at improving diagnostic precision and serving as a reliable tool to physicians in the early diagnosis of cancer.

Image representation of vibration signals and its application on fault diagnosis of rotating machinery with convolutional neural network: A benchmark study

Intelligent fault diagnosis methods on the basis of two-dimensional (2D) image representation of vibration signals (IRVS) and the convolutional neural network (CNN) have been extensively applied in rotating machinery. However, as a large number of IRVS methods are being used, their performance has not been fairly and comprehensively evaluated, which leads to difficulties for researchers to make choices. To address this issue, this paper conducted a benchmark study aimed at comparing the comprehensive performance of different IRVS methods, including computation time, fault diagnosis accuracy, and noise resistance performance. Firstly, the IRVS methods and the fault diagnosis method combining IRVS and CNN were summarized. Then, the general process of the IRVS-CNN method was proposed. Finally, 17 types of IRVS methods were selected, and the performance of the IRVS method was compared based on three datasets and two classic CNN models. The results indicate that the computing time taken to generate 2D images using unthresholded recurrence plot (UTRP), and frequency-domain unthresholded recurrence plot (FDUTRP) methods is relatively longer. The vast majority of frequency-domain methods and time-frequency domain methods can achieve or approach a high accuracy rate in all experiments, demonstrating their outstanding performance. Frequency-domain color vibration image (FDCVI), FDUTRP, and continuous wavelet transform (CWT) have good noise resistance. These results provide a reference for future researchers in selecting IRVS methods.

PointCloud-At: Point Cloud Convolutional Neural Networks with Attention for 3D Data Processing

The rapid growth in technologies for 3D sensors has made point cloud data increasingly available in different applications such as autonomous driving, robotics, and virtual and augmented reality. This raises a growing need for deep learning methods to process the data. Point clouds are difficult to be used directly as inputs in several deep learning techniques. The difficulty is raised by the unstructured and unordered nature of the point cloud data. So, machine learning models built for images or videos cannot be used directly on point cloud data. Although the research in the field of point clouds has gained high attention and different methods have been developed over the decade, very few research works directly with point cloud data, and most of them convert the point cloud data into 2D images or voxels by performing some pre-processing that causes information loss. Methods that directly work on point clouds are in the early stage and this affects the performance and accuracy of the models. Advanced techniques in classical convolutional neural networks, such as the attention mechanism, need to be transferred to the methods directly working with point clouds. In this research, an attention mechanism is proposed to be added to deep convolutional neural networks that process point clouds directly. The attention module was proposed based on specific pooling operations which are designed to be applied directly to point clouds to extract vital information from the point clouds. Segmentation of the ShapeNet dataset was performed to evaluate the method. The mean intersection over union (mIoU) score of the proposed framework was increased after applying the attention method compared to a base state-of-the-art framework that does not have the attention mechanism.

EffSVMNet: An efficient hybrid neural network for improved skin disease classification

The Human Body’s primary defense layer is the skin which protects important organs from various external assaults. This organ protects our internal systems, safeguarding them from possible injury caused by viruses, fungus, and other factors. Unfortunately, the skin is not impenetrable, and infections or damage can occur, which leads to serious problems of health. Even a little skin lesion has the power to become a huge issue. As a result, in our study, our target is to produce an effective system for the quick and early identification of skin illnesses using well-known Convolutional Neural Networks (CNNs). The idea is to use this specialized neural network architecture to improve and speed up the detection and classification process to reduce time-lagging for treatment options. The proposed model i.e., EffSVMNet is a hybrid model consisting of a CNN classifier similar to EfficientNet B3 architecture coupled with a support vector machine (SVM). The sample dataset containing four classes i.e., acne, atopic dermatitis, bullous disease, and eczema is a subset of the DermNet dataset. The proposed model is not only lightweight but also achieves better validation accuracy when compared to similar methods in its category.

Time-guided convolutional neural networks for spatiotemporal urban flood modelling

Urban flood modelling is key to understand flood risks and develop effective interventions in flood management. Deep learning (DL), known for its robust and automatic feature extraction capabilities, has been applied for urban flood predictions. However, the hybrid spatiotemporal structure of conventional DL-enabled urban flood models is limited in terms of accuracy and efficiency. To address this gap, this study develops a new DL model guided by time information. This model uses a classic CNN (Convolution Neural Network) architecture, Unet, as its backbone. Time information is integrated into inputs via an extra channel to specify the desired prediction time, facilitating the simulation of the spatiotemporal flood process. Additionally, a modified loss function is formulated to tackle the sample imbalance problem between flooded and non-flooded sites. The model performance is assessed in an urban area in Dalian, China with a total of 18 rainfall events of varying return periods. The model attains average precision and recall values of 0.90 and 0.81, respectively, across different time steps during various events. Furthermore, the model exhibits transferability in ungauged regions where a high influence of surrounding environments on local flood processes is identified by Grad-CAM (Gradient-weighted Class Activation Mapping) analysis. The results show that the new Unet model has great promise in efficiently providing accurate spatiotemporal flood simulations. The time-guided Unet model can serve as practical tools for rapid flood simulation in urban areas.

ANN uncertainty estimates in assessing fatty liver content from ultrasound data

Background and objectiveThis article uses three different probabilistic convolutional architectures applied to ultrasound image analysis for grading Fatty Liver Content (FLC) in Metabolic Dysfunction Associated Steatotic Liver Disease (MASLD) patients. Steatosis is a new silent epidemic and its accurate measurement is an impelling clinical need, not only for hepatologists, but also for experts in metabolic and cardiovascular diseases. This paper aims to provide a robust comparison between different uncertainty quantification strategies to identify advantages and drawbacks in a real clinical setting. MethodsWe used a classical Convolutional Neural Network, a Monte Carlo Dropout, and a Bayesian Convolutional Neural Network with the goal of not only comparing the goodness of the predictions, but also to have access to an evaluation of the uncertainty associated with the outputs. ResultsWe found that even if the prediction based on a single ultrasound view is reliable (relative RMSE [5.93%-12.04%]), networks based on two ultrasound views outperform them (relative RMSE [5.35%-5.87%]). In addition, the results show that the introduction of a “not confident” category contributes to increase the percentage of correctly predicted cases and to decrease the percentage of mispredicted cases, especially for semi-intrusive methods. ConclusionsThe possibility of having access to information about the confidence with which the network produces its outputs is a great advantage, both from the point of view of physicians who want to use neural networks as computer-aided diagnosis, and for developers who want to limit overfitting and obtain information about dataset problems in terms of out-of-distribution detection.

Advanced Dual-Input Artificial Optical Synapse for Recognition and Generative Neural Network

Perovskite materials have emerged as leading candidates for optical synaptic devices due to their superior photosensitivity and tunable optoelectronic properties. However, the practical application of perovskite-based optoelectronic synaptic transistors has been hindered by issues of poor stability and high toxicity. This study developed an artificial synaptic thin film transistor (TFT) based on a Cs2AgBiBr6/InOx bilayer. The device demonstrated synaptic behavior under optoelectronic hybrid stimulation (largest wavelength ~520nm), such as excitatory postsynaptic current (EPSC), inhibitory postsynaptic current (IPSC), paired-pulse facilitation (PPF), spike-timing-dependent plasticity (STDP), short-term memory (STM) and long-term memory (LTM). Notably, the “light writing and voltage erasing” characteristics of these devices could be utilized to construct a convolutional neural network (CNN) classifier for the CIFAR-10 dataset, demonstrating noise tolerance close to the human eye. The loss in recognition accuracy was within 1% when Gaussian white noise and salt pepper noise were added. Furthermore, these devices exhibited great potential in Generative Adversarial Networks (CycleGAN), with the generated image quality achieving levels of 0.637 and 0.715 in Improved Perceptual Image Processing System (IPIPS) and Structural Similarity Index (SSIM) evaluation metrics for the zebra dataset, indicating good image quality. This study indicates that our Cs2AgBiBr6-based artificial optical synaptic thin-film transistors (TFTs) are promising for sensing and in-memory computing applications.

Fractional Gradient Optimized Explainable Convolutional Neural Network for Alzheimer's Disease Diagnosis

Alzheimer’s is one of the brain syndromes that steadily effects the brain memory. The early stage of Alzheimer’s disease referred as mild-cognitive-impairment (MCI) and the growth of Alzheimer’s is not certain in the patients with MCI. The premature Alzheimer’s diagnosis detection is crucial for the healthy brains functionality by avoiding memory loss. Different multi-neural network architectures have been suggested by the researchers for efficient and accurate Alzheimer’s disease (AD) detection. The absence of improved feature extraction mechanisms and unexplored efficient optimizers in complex benchmark architectures lead to an inefficient and inaccurate AD classification. Moreover, the standard Convolutional Neural Network (CNN)-based architectures for Alzheimer’s diagnosis lack interpretability in their predictions. An interpretable, simplified yet effective deep learning model is required for accurate classification of Alzheimer’s disease. In this study, a generalized fractional order-based CNN classifier with explainable artificial intelligence (XAI) capabilities is proposed for accurate, efficient and interpretable classification of AD diagnosis. The proposed study (a) classifies AD accurately by incorporating unexplored pooling technique with enhanced feature extraction mechanism (b) provides fractional order-based optimization approach for adaptive learning and fast convergence speed (c) suggests an interpretable method for proving the transparency of the model. The proposed model outclasses complex benchmark architectures with regard to accuracy using standard ADNI dataset. The suggested fractional order-based CNN classifier achieves an improved accuracy of 99% as compared to the state-of-the-art models.

Tumor detection in breast cancer pathology patches using a Multi-scale Multi-head Self-attention Ensemble Network on Whole Slide Images

Breast cancer (BC) is the most common type of cancer among women globally and is one of the leading causes of cancer-related deaths among women. In the diagnosis of BC, histopathological assessment is the gold standard, where automated tumor detection technologies play a pivotal role. Utilizing Convolutional Neural Networks (CNNs) for automated analysis of image patches from Whole Slide Images (WSIs) enhances detection accuracy and alleviates the workload of pathologists. However, CNNs often face limitations in handling pathological patches due to a lack of sufficient contextual information and limited feature generation capabilities. To address this, we propose a novel Multi-scale Multi-head Self-attention Ensemble Network (MMSEN), which integrates a multi-scale feature generation module, a convolutional self-attention module, and an adaptive feature integration and output module, effectively optimizing the performance of classical CNNs. The design of MMSEN optimizes the capture of key information and the comprehensive integration of features in WSIs pathological patches, significantly enhancing the precision of tumor detection. Validation results from a five-fold cross-validation experiment on the PatchCamelyon (PCam) dataset demonstrate that MMSEN achieves a ROC-AUC of 99.01%±0.02%, an F1 Score of 98.00%±0.08%, a Balanced Accuracy (B-Acc) of 98.00%±0.08%, and a Matthews Correlation Coefficient (MCC) of 96.00%±0.16% (p<0.05). These results demonstrate the effectiveness and potential of MMSEN in detecting tumors from pathological patches in WSIs for BC.

A convolutional neural network model detecting lasting behavioral changes in mice with kanamycin-induced unilateral inner ear dysfunction

In acute aminoglycoside ototoxicity of the unilateral inner ear, physical abnormalities, such as nystagmus and postural alteration, are relieved within a few days by neural compensation. To examine exploratory behavior over an extended period, behaviors of freely moving mice after unilateral kanamycin injection into the inner ear were recorded in a home cage environment. The tail was excluded from deep learning-mediated object detection because of its delayed movement relative to the body. All detection results were confirmed using a convolutional neural network classification model. In kanamycin-injected mice, the total distance moved in 15 min increased on postoperative day 3. Furthermore, injured mice turned more frequently toward the healthy side up to 17 days after the surgery. This tendency resulted in increased clockwise movements in home cage recordings. Moreover, tail suspension and twisting toward the healthy side induced a physical sign for up to 14 days after the injury; the mice rapidly rotated with dorsal bending. Our analysis strategy employing deep learning helps to evaluate neuronal compensatory processes for an extended period and is useful for assessing the efficacy of therapeutic interventions.

Indonesian Fake News Classification Using Transfer Learning in CNN and LSTM

Fake news spreads quickly and is challenging to stop due to the ease of accessing and sharing information online. Deep learning techniques are a method that can be used to identify fake news quickly and accurately. The types of neural networks commonly utilized in deep learning architectures include Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM), which can perform well when managing the task of classifying fake news, according to several pertinent studies. Regarding handling instances of Indonesian fake news classification, this study compares how well the CNN and LSTM models perform. However, given that Indonesian is a low-resource language with scant documentation, it is challenging to build an adequate data set. At the same time, the CNN and LSTM classification models require significant training data. We proposed a transfer learning method by combining two classification models with a pre-trained IndoBERT language model. 1340 news text data were used, including 643 actual news texts from CNN Indonesia, Liputan6, and Detik and 697 fake news texts from TurnBackHoax. As a result, the performance of the combination of the LSTM classification model with IndoBERT outperformed that of the CNN classification model with IndoBERT, which only produced an accuracy of 92.91%, down by 6%, and was able to produce an accuracy of up to 97.76%, an increase of 4.8% from before. Furthermore, the results show that the LSTM classification model outperforms the CNN classification model in capturing the representation created by IndoBERT. Additionally, these insights may serve as a basis for future research on identifying fake news in Indonesia, helping to improve methods for combatting misinformation in Indonesia.