Deep Learning-based Multi-class Three-dimensional (3-D) Object Classification using Phase-only Digital Holographic Information

BackgroundAntimicrobial peptides (AMPs) are short peptides with diverse biological activities and playing a crucial role in various biological processes. Due to the widespread misuse of traditional antibiotics and the increasing resistance of microorganisms to these drugs, AMPs have emerged as a promising alternative. Consequently, the identification of AMPs has garnered significant research interest. Numerous computational methods based on machine learning algorithms have been developed to facilitate AMP recognition. However, some existing AMPs recognition models only focus on binary classification tasks or only identify the functional activity of a limited number of AMPs categories in multi-class classification tasks. To address this limitation, this study proposes a two-stage AMPs recognition model, iAMP-SeE.MethodsThe iAMP-SeE model extracts features from protein sequences using ESM2, employs a Convolutional Neural Network (CNN) module to capture local patterns from ESM features and utilizes a Bidirectional Long Short-Term Memory (BiLSTM) network to capture long-term dependencies. Furthermore, it incorporates Squeeze-and-Excitation (SE) and Efficient Channel Attention (ECA) mechanisms, which focus on global and local channel relationships, respectively. These two attention mechanisms are complementary, as they enhance features across various dimensions and granularities while simultaneously suppressing irrelevant or redundant features, thereby boosting the model’s performance. Additionally, to address the issue of imbalanced datasets, the Synthetic Minority Over-sampling Technique (SMOTE) is incorporated into the multi-classification task. This method balances the number of AMP categories and ensures that minority classes are not overlooked during model training.ResultsEvaluation across a range of classification thresholds demonstrated the stability of the model’s performance metrics in both binary and multi-class tasks. Furthermore, comparative experiments with existing AMP recognition models confirmed the superior performance of iAMP-SeE.ConclusionsRigorous experimental comparisons and ablation studies demonstrate the effectiveness of iAMP-SeE for both binary and multi-class AMP classification tasks. The source code is publicly available at: https://github.com/cqw0715/iAMP-SeE.git.

Wearable electroencephalography (EEG) devices for seizure detection accuracy and reliability are deep learning (DL) applications in the field of epilepsy diagnosis. In this study, we sought to increase the accuracy of seizure detection using advanced DL algorithms on the Children’s Hospital Boston - Massachusetts Institute of Technology (CHB-MIT) EEG database. First, a fully convolutional network (FCN) was trained and assessed using accuracy and recall/precision metrics, and the early stopping technique was used to avoid overfitting. To assess the performance, the FCN was evaluated in terms of various metrics, including accuracy, precision, recall, F1-score, and receiver operating characteristic (ROC)-area under the curve (AUC). In addition, two-dimensional (2D) convolutional neural networks (CNNs) and long short-term memory (LSTM) models were used to model the database, and their performance was thoroughly measured using different metrics, graphs, and confusion matrices. Using LSTM variants, such as standard LSTM, bidirectional LSTM, stacked LSTM, and LSTM attention mechanisms, hybrid convolutional LSTM (ConvLSTM) models were trained and compared. The comparison was conducted based on the training and validation accuracy and loss, as well as the graphs resulting from the precision–recall curves. Apart from DL approaches, EEG signal analysis using time–frequency techniques, such as wavelet transform and short-time Fourier transform, has also been investigated. These methods assisted in the analysis of the time–frequency features of EEG signals in combination with DL models. This study demonstrates that the performance of wearable EEG devices can be augmented using a combination of DL and seizure signal processing techniques. The FCN achieved an accuracy of 92%, with a recall for seizures of 33%, an F1-score of 0.03, and strong ROC-AUC results. The 2D CNN achieved 96% accuracy, a seizure recall of 70%, an F1-score of 0.12, and an ROC-AUC score of 78%. The baseline LSTM struggled with effectiveness at 53% accuracy with a seizure recall of 18%. In contrast, the LSTM model, which incorporated synthetic minority oversampling technique (SMOTE) balancing, was able to reach improvements of up to 89% accuracy, with a precision of 91%, a recall of 86%, an F1-score of 0.89, and a strong ROC curve performance. Among the models, the LSTM with SMOTE was the best performer, with 89% accuracy, 91% precision, 86% recall, and an F1-score of 0.89. These results provide evidence that applying techniques for data balancing in combination with certain DL network architectures significantly improves the detection of seizures using wearable EEG devices worn on the body. We believe that real-time monitoring and high-performance systems are feasible using optimized DL frameworks. The analysis of the performance of different models allows for the understanding of the possibilities of optimizing the architectures of DL algorithms for the modern diagnosis of epilepsy in real time. The source code used to carry out the experiments is publicly available at CHB-MIT EEG Dataset Python Scripts (https://www.kaggle.com/code/adnankust/adnaneeg1).

Convolutional Neural Networks (CNNs) demonstrate state-of-the-art performance in large-scale multi-class image classification tasks. CNNs consist of convolution layers that progressively construct features and a classification layer. Typically, a softmax function is used in the classification layer to learn joint probabilities for the classes, which are subsequently used for class prediction. We refer to such an approach as the joint approach to multi-class classification. There exists another approach in the literature which determines the multi-class prediction outcome through a sequence of binary decisions, and is christened the class binarization approach. A popular type of class binarization is error-correcting output codes. In this paper, we propose to incorporate error-correcting output codes into convolutional neural networks by inserting a latent binarization layer in a CNN classification layer. This approach encapsulates both encoding and decoding steps of error-correcting output codes into a single CNN architecture that is capable of discovering an optimal coding matrix during training. The latent binarization layer is motivated by the family of latent-trait and latent-class models used in behavioral research. We call the proposed convolutional neural networks with Latent Binarization as LB-CNNs, and develop algorithms combining expectation maximization algorithms and back-propagation to train LB-CNNs. The proposed models and algorithms are applied to several image recognition tasks, producing excellent results. Furthermore, our model enhances the interpretability of the decision process of standard convolutional neural networks.

Strabismus, or eye misalignment, is a common condition affecting individuals of all ages. Early detection and accurate classification are essential for proper treatment and avoiding long-term complications. This research presents a new deep-learning-based approach for automatically identifying and classifying strabismus from facial images. The proposed methodology leverages Convolutional Neural Networks (CNNs) to achieve high accuracy in both binary (strabismus vs. normal) and multi-class (eight-class deviation angle for esotropia and exotropia) classification tasks. The dataset for binary classification consisted of 4,257 facial images, including 1,599 normal cases and 2,658 strabismus cases, while the multi-class classification involved 480 strabismic and 142 non-strabismic images. These images were labeled based on ophthalmologist measurements using the Alternate Prism Cover Test (APCT) or the Modified Krimsky Test (MK). Five-fold cross-validation was employed, and performance was evaluated using sensitivity, accuracy, F1-score, and recall metrics. The proposed deep learning model achieved an accuracy of 86.38% for binary classification and 92.7% for multi-class classification. These results demonstrate the potential of our approach to assist healthcare professionals in early strabismus detection and treatment planning, ultimately improving patient outcomes.

BackgroundTranscutaneous transhepatic gallbladder drainage (PTGBD) has shown significant efficacy in the treatment of elderly patients with acute cholecystitis. The goal of this study is to develop a machine learning-based web calculator aimed at predicting the optimal timing for cholecystectomy (LC) after PTGBD in elderly patients with acute cholecystitis (AC) to achieve precise personalized medicine.MethodsA retrospective analysis of 979 elderly patients with acute cholecystitis admitted to Jinzhou Central Hospital and the First Affiliated Hospital of Jinzhou Medical University from 2013 to 2024 was performed, and a total of 680 patients were included in the model development. Patients were divided into delayed (347 cases, surgery > 6 weeks post-PTGBD) and non-delayed (333 cases) groups based on the interval between PTGBD and LC. Minimal Absolute Contraction and Selection Operator (LASSO) and logistic analysis were used to determine the predictors of postponement of LC in elderly patients with AC after PTGBD. Next, we used eight ML algorithms, namely Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), Extreme gradient boosting (XGB), Support Vector Machine (SVM), Multilayer Perceptron (MLP), K-nearest Neighbor (KNN), Gaussian Naive Bayes (GNB), to train and develop ML models using a 10x cross-validation method. The performance of the model was evaluated by a variety of indicators, including the area under the receiver operating characteristic curve (ROC), calibration curve, decision curve, PR curve, and confusion matrix. In addition, model interpretation is performed through Shapley Additive Interpretation (SHAP) analysis to clarify the importance of each feature of the model and its basis for decision-making. Finally, we chose to use the best model to develop a web-based calculator that could be used to predict the likelihood of delaying LC after PTGBD in elderly AC patients.ResultsIn multivariate logistic regression analysis, age, sex, gallbladder wall thickness, time between onset and PTGBD, white blood cell count (WBC), C-reactive protein (CRP), and neutrophil-to-lymphocyte ratio (NLR) were identified as independent predictors of delayed LC in elderly patients with AC after PTGBD. In the training set, the area under the receiver operating characteristic curve (AUC) values for these models ranged from 0.808 to 0.914, with the random forest (RF) model showing the highest AUC value. Through the evaluation of decision curve analysis (DCA), precision-recall (PR) curve and calibration curve, the RF model showed superior clinical decision support and prediction performance compared with the other seven models. Finally, we used the RF model to build an online network calculator, which aims to accurately assist doctors in making more informed and accurate clinical decisions and promote the wide application of the model in clinical practice (https://zw17786325639.shinyapps.io/Postpone/).ConclusionsThis study developed and validated an RF model network calculator based on clinical indicator information to assess the likelihood of delaying LC after PTGBD in elderly patients with acute cholecystitis. This tool is expected to assist physicians in making more appropriate clinical decisions for patients.

The research presented in this dissertation represents an innovative application of computer‐aided diagnosis and signal detection theory to the specific task of early detection of breast cancer in the context of screening mammography. A number of automated schemes have been developed in our laboratory to detect masses and clustered microcalcifications in digitized mammograms, on the one hand, and to classify known lesions as malignant or benign, on the other. The development of fully automated classification schemes is difficult, because the output of a detection scheme will contain false‐positive detections in addition to detected malignant and benign lesions, resulting in a three‐class classification task. Researchers have so far been unable to extend successful tools for analyzing two‐class classification tasks, such as receiver operating characteristic (ROC) analysis, to three‐class classification tasks. The goals of our research were to use Bayesian artificial neural networks to estimate ideal observer decision variables to both detect and classify clustered microcalcifications and mass lesions in mammograms, and to derive substantial theoretical results indicating potential avenues of approach toward the three‐class classification task. Specifically, we have shown that an ideal observer in an N‐class classification task achieves an optimal ROC hypersurface, just as the two‐class ideal observer achieves an optimal ROC curve; and that an obvious generalization of a well‐known two‐class performance metric, the area under the ROC curve, is not useful as a performance metric in classification tasks with more than two classes. This work is significant for three reasons. First, it involves the explicit estimation of feature‐based (as opposed to image‐based) ideal observer decision variables in the tasks of detecting and classifying mammographic lesions. Second, it directly addresses the three‐class classification task of distinguishing malignant lesions, benign lesions, and false‐positive computer detections. Finally, it develops important theoretical results for N‐class classification tasks that should prove of value in the development of a three‐class extension to ROC analysis methods.

Despite the significant progress in diagnosis and treatment, breast cancer still continues to be the most common deadly disease occurring among women in the world. Histopathological analysis is the standard method for breast cancer diagnosis. Manual inspection of histological samples is a highly specialized and time consuming task. Computer-Aided Diagnosis Systems (CAD) could overcome the limitations of manual analysis and assist pathologists to diagnose cancer more accurately and efficiently. Hence, we propose a novel approach that uses deep learned nucleus feature classifier fusion framework for breast cancer detection from histopathological images. Recently, convolutional neural networks (CNNs) have produced high recognition rates compared to hand crafted features but results in much higher complexity in network design. Moreover, designing CNN from scratch requires large amount of training data. So, we use pretrained CNNs for feature extraction. Extracted features are fused using norm pooling and fed to Support Vector Machine (SVM) for classification. Classification results using the different CNNs are fused using belief theory based fusion. Performance analysis is done using accuracy, sensitivity, specificity, receiver operating characteristics (ROC) curve and AUC (Area under the ROC curve) measures. Results comparable to the current state of the art is achieved.

Alzheimer’s Disease (AD) is a health apprehension of significant proportions that is negatively impacting the ageing population globally. It is characterized by neuronal loss and the formation of structures such as neurofibrillary tangles and amyloid plaques in the early as well as later stages of the disease. Neuroimaging modalities are routinely used in clinical practice to capture brain alterations associated with AD. On the other hand, deep learning methods are routinely used to recognize patterns in underlying data distributions effectively. This work uses Convolutional Neural Network (CNN) architectures in both 2D and 3D domains to classify the initial stages of AD into AD, Mild Cognitive Impairment (MCI) and Normal Control (NC) classes using the positron emission tomography neuroimaging modality deploying data augmentation in a random zoomed in/out scheme. We used novel concepts such as the blurring before subsampling principle and distant domain transfer learning to build 2D CNN architectures. We performed three binaries, that is, AD/NC, AD/MCI, MCI/NC and one multiclass classification task AD/NC/MCI. The statistical comparison revealed that 3D-CNN architecture performed the best achieving an accuracy of 89.21% on AD/NC, 71.70% on AD/MCI, 62.25% on NC/MCI and 59.73% on AD/NC/MCI classification tasks using a five-fold cross-validation hyperparameter selection approach. Data augmentation helps in achieving superior performance on the multiclass classification task. The obtained results support the application of deep learning models towards early recognition of AD.

The paper presents a solution to the multiclass classification problem based on the Convolutional Fuzzy Neural Networks. The proposed model includes a fuzzy self-organization layer for data clustering (in addition to convolutional, pooling and fully-connected layers). The model combines the power of convolutional neural networks and fuzzy logic, it is capable of handling uncertainty and imprecision in the input pattern representation. The Convolutional Fuzzy Neural Networks could provide better accuracy in multiclass classification tasks when classified objects are often characterized by uncertainty and inaccuracy in their representation.

Summary: Precision-recall (PR) and receiver operating characteristic (ROC) curves are valuable measures of classifier performance. Here, we present the R-package PRROC, which allows for computing and visualizing both PR and ROC curves. In contrast to available R-packages, PRROC allows for computing PR and ROC curves and areas under these curves for soft-labeled data using a continuous interpolation between the points of PR curves. In addition, PRROC provides a generic plot function for generating publication-quality graphics of PR and ROC curves.Availability and implementation: PRROC is available from CRAN and is licensed under GPL 3.Contact: grau@informatik.uni-halle.de

Deep learning can discover intricate patterns hidden in big data, and has much better scalability than traditional machine learning when the volume of data increases dramatically. Thus, deep learning has gained many successes in various domains and applications such as image classification, text classification, and machine translation. In this paper, we use deep learning to classify geographical features (e.g., mountains, rivers, landmarks, and cities) from text, using geolocated Wikipedia entries as the case study application. We employ one of the most commonly used deep learning architectures, convolutional neural networks (CNNs) and its integration with active learning (creating what we call active CNNs), to train the geographical feature classifiers on the Wikipedia text data set obtained from GeoNames (which provides the feature type for each geolocated entity). We evaluate the performance of CNNs and active CNNs with multiple metrics (i.e., accuracy, F1 score, and confusion matrix). Our experiment results demonstrated that CNNs and active CNNs can effectively classify geo-referenced text entities into predefined geographical features. In addition, our experiment results show that active CNNs outperform CNNs for hard to distinguish classes. In our experiment, we also compared results for hierarchical multi-class classification and flat multiclass classification, and the results show that hierarchical multiclass classification significantly outperforms flat multi-class classification for the data set we used.

BackgroundKnowledge of catalytic residues can play an essential role in elucidating mechanistic details of an enzyme. However, experimental identification of catalytic residues is a tedious and time-consuming task, which can be expedited by computational predictions. Despite significant development in active-site prediction methods, one of the remaining issues is ranked positions of putative catalytic residues among all ranked residues. In order to improve ranking of catalytic residues and their prediction accuracy, we have developed a meta-approach based method CSmetaPred. In this approach, residues are ranked based on the mean of normalized residue scores derived from four well-known catalytic residue predictors. The mean residue score of CSmetaPred is combined with predicted pocket information to improve prediction performance in meta-predictor, CSmetaPred_poc.ResultsBoth meta-predictors are evaluated on two comprehensive benchmark datasets and three legacy datasets using Receiver Operating Characteristic (ROC) and Precision Recall (PR) curves. The visual and quantitative analysis of ROC and PR curves shows that meta-predictors outperform their constituent methods and CSmetaPred_poc is the best of evaluated methods. For instance, on CSAMAC dataset CSmetaPred_poc (CSmetaPred) achieves highest Mean Average Specificity (MAS), a scalar measure for ROC curve, of 0.97 (0.96). Importantly, median predicted rank of catalytic residues is the lowest (best) for CSmetaPred_poc. Considering residues ranked ≤20 classified as true positive in binary classification, CSmetaPred_poc achieves prediction accuracy of 0.94 on CSAMAC dataset. Moreover, on the same dataset CSmetaPred_poc predicts all catalytic residues within top 20 ranks for ~73% of enzymes. Furthermore, benchmarking of prediction on comparative modelled structures showed that models result in better prediction than only sequence based predictions. These analyses suggest that CSmetaPred_poc is able to rank putative catalytic residues at lower (better) ranked positions, which can facilitate and expedite their experimental characterization.ConclusionsThe benchmarking studies showed that employing meta-approach in combining residue-level scores derived from well-known catalytic residue predictors can improve prediction accuracy as well as provide improved ranked positions of known catalytic residues. Hence, such predictions can assist experimentalist to prioritize residues for mutational studies in their efforts to characterize catalytic residues. Both meta-predictors are available as webserver at: http://14.139.227.206/csmetapred/.

This study aims to develop an attention-based deep learning model for distinguishing benign from malignant breast lesions on CESM. Preoperative CESM images of 1239 patients, which were definitely diagnosed on pathology in a multicentre cohort, were divided into training and validation sets, internal and external test sets. The regions of interest of the breast lesions were outlined manually by a senior radiologist. We adopted three conventional convolutional neural networks (CNNs), namely, DenseNet 121, Xception, and ResNet 50, as the backbone architectures and incorporated the convolutional block attention module (CBAM) into them for classification. The performance of the models was analysed in terms of the receiver operating characteristic (ROC) curve, accuracy, the positive predictive value (PPV), the negative predictive value (NPV), the F1 score, the precision recall curve (PRC), and heat maps. The final models were compared with the diagnostic performance of conventional CNNs, radiomics models, and two radiologists with specialised breast imaging experience. The best-performing deep learning model, that is, the CBAM-based Xception, achieved an area under the ROC curve (AUC) of 0.970, a sensitivity of 0.848, a specificity of 1.000, and an accuracy of 0.891 on the external test set, which was higher than those of other CNNs, radiomics models, and radiologists. The PRC and the heat maps also indicated the favourable predictive performance of the attention-based CNN model. The diagnostic performance of two radiologists improved with deep learning assistance. Using an attention-based deep learning model based on CESM images can help to distinguishing benign from malignant breast lesions, and the diagnostic performance of radiologists improved with deep learning assistance.

The use of convolutional neural networks (CNNs) in the medical field has revolutionized diagnosis, particularly for prostate cancer, one of the most common cancers in men. Early diagnosis is crucial to improve treatment outcomes. MRI is a key imaging modality, providing detailed information about prostate tissue, but its interpretation can be subject to subjectivity depending on the expertise of radiologists. The application of CNNs for prostate MRI analysis represents a significant advance in diagnosis. This model can thus serve as a decision support tool for radiologists, improving diagnostic accuracy and, potentially, clinical outcomes. The proposed model demonstrated high predictive power for prostate cancer, with high performance in terms of accuracy, sensitivity, and specificity. ROC curves illustrated good discrimination between positive and negative cases. This approach demonstrates the potential of artificial intelligence in complex medical diagnosis, and future studies could explore the integration of this model into clinical systems for use in real-world settings. The proposed model for prostate cancer prediction demonstrated excellent performance, particularly in terms of accuracy, sensitivity, and specificity. The model's effectiveness was evaluated using Receiver Operating Characteristic (ROC) curves, a fundamental tool for assessing the performance of binary classifiers. These curves plot the true positive rate (sensitivity) against the false positive rate, allowing the model's ability to correctly distinguish between positive cases (patients with cancer) and negative cases (healthy patients) to be evaluated. The closer the ROC curve is to the top-left corner of the plot, the better the model's performance. In this study, the resulting ROC curve exhibited outstanding discrimination, confirming the robustness of the proposed predictive system. These findings highlight the potential of artificial intelligence, particularly neural networks, in the field of computer-aided medical diagnosis. Future research could focus on the practical integration of this model into clinical systems to enhance early detection of prostate cancer in real-world settings.

We argue that citations, as they have different reasons and functions, should not all be treated in the same way. Using the large, annotated dataset of about 10K citation contexts annotated by human experts, extracted from the Association for Computational Linguistics repository, we present a deep learning–based citation context classification architecture. Unlike all existing state-of-the-art feature-based citation classification models, our proposed convolutional neural network (CNN) with fastText-based pre-trained embedding vectors uses only the citation context as its input to outperform them in both binary- (important and non-important) and multi-class (Use, Extends, CompareOrContrast, Motivation, Background, Other) citation classification tasks. Furthermore, we propose using focal-loss and class-weight functions in the CNN model to overcome the inherited class imbalance issues in citation classification datasets. We show that using the focal-loss function with CNN adds a factor of [Formula: see text] to the cross-entropy function. Our model improves on the baseline results by achieving an encouraging 90.6 F1 score with 90.7% accuracy and a 72.3 F1 score with a 72.1% accuracy score, respectively, for binary- and multi-class citation classification tasks.