Neural Network-based Classification Research Articles

A position-enhanced sequential feature encoding model for lung infections and lymphoma classification on CT images.

Differentiating pulmonary lymphoma from lung infections using CT images is challenging. Existing deep neural network-based lung CT classification models rely on 2D slices, lacking comprehensive information and requiring manual selection. 3D models that involve chunking compromise image information and struggle with parameter reduction, limiting performance. These limitations must be addressed to improve accuracy and practicality. We propose a transformer sequential feature encoding structure to integrate multi-level information from complete CT images, inspired by the clinical practice of using a sequence of cross-sectional slices for diagnosis. We incorporate position encoding and cross-level long-range information fusion modules into the feature extraction CNN network for cross-sectional slices, ensuring high-precision feature extraction. We conducted comprehensive experiments on a dataset of 124 patients, with respective sizes of 64, 20 and 40 for training, validation and testing. The results of ablation experiments and comparative experiments demonstrated the effectiveness of our approach. Our method outperforms existing state-of-the-art methods in the 3D CT image classification problem of distinguishing between lung infections and pulmonary lymphoma, achieving an accuracy of 0.875, AUC of 0.953 and F1 score of 0.889. The experiments verified that our proposed position-enhanced transformer-based sequential feature encoding model is capable of effectively performing high-precision feature extraction and contextual feature fusion in the lungs. It enhances the ability of a standalone CNN network or transformer to extract features, thereby improving the classification performance. The source code is accessible at https://github.com/imchuyu/PTSFE .

Unbiased analysis of spatial learning strategies in a modified Barnes maze using convolutional neural networks

Assessment of spatial learning abilities is central to behavioral neuroscience and a useful tool for animal model validation and drug development. However, biases introduced by the apparatus, environment, or experimentalist represent a critical challenge to the test validity. We have recently developed the Modified Barnes Maze (MBM) task, a spatial learning paradigm that overcomes inherent behavioral biases of animals in the classical Barnes maze. The specific combination of spatial strategies employed by mice is often considered representative of the level of cognitive resources used. Herein, we have developed a convolutional neural network-based classifier of exploration strategies in the MBM that can effectively provide researchers with enhanced insights into cognitive traits in mice. Following validation, we compared the learning performance of female and male C57BL/6J mice, as well as that of Ts65Dn mice, a model of Down syndrome, and 5xFAD mice, a model of Alzheimer’s disease. Male mice exhibited more effective navigation abilities than female mice, reflected in higher utilization of effective spatial search strategies. Compared to wildtype controls, Ts65Dn mice exhibited delayed usage of spatial strategies despite similar success rates in completing this spatial task. 5xFAD mice showed increased usage of non-spatial strategies such as Circling that corresponded to higher latency to reach the target and lower success rate. These data exemplify the need for deeper strategy classification tools in dissecting complex cognitive traits. In sum, we provide a machine-learning-based strategy classifier that extends our understanding of mice’s spatial learning capabilities while enabling a more accurate cognitive assessment.

Three-dimensional convolutional neural network-based classification of chronic kidney disease severity using kidney MRI

A three-dimensional convolutional neural network model was developed to classify the severity of chronic kidney disease (CKD) using magnetic resonance imaging (MRI) Dixon-based T1-weighted in-phase (IP)/opposed-phase (OP)/water-only (WO) imaging. Seventy-three patients with severe renal dysfunction (estimated glomerular filtration rate [eGFR] < 30 mL/min/1.73 m2, CKD stage G4–5); 172 with moderate renal dysfunction (30 ≤ eGFR < 60 mL/min/1.73 m2, CKD stage G3a/b); and 76 with mild renal dysfunction (eGFR ≥ 60 mL/min/1.73 m2, CKD stage G1–2) participated in this study. The model was applied to the right, left, and both kidneys, as well as to each imaging method (T1-weighted IP/OP/WO images). The best performance was obtained when using bilateral kidneys and IP images, with an accuracy of 0.862 ± 0.036. The overall accuracy was better for the bilateral kidney models than for the unilateral kidney models. Our deep learning approach using kidney MRI can be applied to classify patients with CKD based on the severity of kidney disease.

Quantum state reconstruction in a noisy environment via deep learning

Quantum noise is currently limiting efficient quantum information processing and computation, impacting on the fidelity and reliability of quantum states. In this work, we consider the tasks of reconstructing and classifying quantum states corrupted by the action of an unknown noisy channel using classical feed-forward neural networks. By framing reconstruction as a regression problem, we show how such an approach can be used to recover with fidelities exceeding 99% the noiseless density matrices of quantum states of up to three qubits undergoing noisy evolution, and we test its performance with both single-qubit (bit-flip, phase-flip, depolarizing, and amplitude damping) and two-qubit quantum channels (correlated amplitude damping). Furthermore, a critical aspect of our investigation involves also a comprehensive comparison between mean squared error and infidelity as loss functions. Our findings reveal that these two metrics yield comparable results in the context of state reconstruction. Moreover, we also consider the task of distinguishing between different quantum noisy channels, and show how a neural network-based classifier is able to solve such a classification problem with perfect accuracy.

A neural network-based classifier for identifying and locating neutral wire breaks in low voltage distribution networks

The breakage of the neutral conductor in low voltage distribution networks is a major concern for distribution companies. This breakage causes significant voltage deviations that can damage the connected equipment as well as jeopardizing people. The detection and localization of the breakage is a major challenge as it does not always manifest in the same way. This work presents a methodology based on artificial intelligence for the detection and localization of neutral conductor breaks in distribution networks. Two neural networks are trained in attempt to solve each of the challenges. For this purpose, measurements commonly taken by smartmeters such as power and nodal voltages are used. The methodology is evaluated in simulation exhibiting a good performance.

Machine-learning guided differentiation between photoplethysmography waveforms of supraventricular and ventricular origin

Abstract Aims To test, whether a neural network-based classifier could aid in distinguishing photoplethysmographic (PPG) pulse waveforms of ventricular from those of supraventricular origin. Methods Thirty patients undergoing invasive electrophysiological studies for documented narrow complex tachycardias between May and December 2021 at a university medical centre were included in the trial. PPG waveforms were recorded using a PPG wristband (Empatica E4) in parallel to 12-lead surface electrograms (ECGs) and intracardiac electrograms from diagnostic catheters. Pulse waves were then annotated to either supraventricular (sinus rhythm, atrial pacing) or ventricular origin (ventricular pacing) based on invasive electrogram recordings, ECGs and stimulation protocols. 21147 data sets were split into test, training and validation data sets and used to develop, train and validate a convolutional neural network (CNN) for classifying PPG waveforms according to their origin into sinus rhythm, atrial or ventricular pacing. Results Datasets were complete for 27 patients. 74% were female and median age was 53 (range 18, 78) years, median BMI was 27±5, 19% had arterial hypertension, 11% had diabetes and 15% had coronary artery disease. During the electrophysiological study, typical AV-nodal re-entrant tachycardias could be induced in 63% of patients, 15% had inducible atrial tachycardias and the other patients had no inducible tachyarrhythmias. Our CNN-based classifier was able to correctly predict the origins of pulse waves in 95±1% for sinus rhythm, 96±1% for atrial paced rhythms and 95±1% for ventricular rhythms. Conclusion A neural network-based classifier trained on ground truth PPG data obtained by using invasive EP measurements for annotation could accurately predict atrial or ventricular origin from PPG waveforms alone.Graphical abstract

Initializing the weights of a multilayer perceptron for activity and emotion recognition

Conducting an analysis of human behavior is an intriguing topic for many researchers. Within this field, machine learning can be applied to classify activities and emotions by analyzing physiological signals. However, the limited size of available databases poses challenges for the generalization of classifiers. This paper proposes a method to enhance the generalization of neural network-based classifiers by intelligently initializing weights for emotion and activity recognition. The signals under consideration are electrocardiogram, thoracic electrical bioimpedance, and electrodermal activity. The database used comprises recordings from 40 subjects performing various tasks that induce emotions and activities. The performance of the proposed method is compared with several standard machine learning and deep learning classifiers typically employed in emotion and activity recognition. This study involves two primary assessments. First is the activity recognition task, encompassing classes such as neutral, emotional, mental, and physical activity, where results close to 20% accuracy are achieved using the three physiological signals. Second, the emotion recognition assessment aims to differentiate between emotions like neutral, sadness, and disgust. An error probability close to 15% is obtained using thoracic electrical bioimpedance and electrodermal activity. The proposed method yields the best results among the approaches evaluated.

A multifaceted suite of metrics for comparative myoelectric prosthesis controller research.

Upper limb robotic (myoelectric) prostheses are technologically advanced, but challenging to use. In response, substantial research is being done to develop person-specific prosthesis controllers that can predict a user's intended movements. Most studies that test and compare new controllers rely on simple assessment measures such as task scores (e.g., number of objects moved across a barrier) or duration-based measures (e.g., overall task completion time). These assessment measures, however, fail to capture valuable details about: the quality of device arm movements; whether these movements match users' intentions; the timing of specific wrist and hand control functions; and users' opinions regarding overall device reliability and controller training requirements. In this work, we present a comprehensive and novel suite of myoelectric prosthesis control evaluation metrics that better facilitates analysis of device movement details-spanning measures of task performance, control characteristics, and user experience. As a case example of their use and research viability, we applied these metrics in real-time control experimentation. Here, eight participants without upper limb impairment compared device control offered by a deep learning-based controller (recurrent convolutional neural network-based classification with transfer learning, or RCNN-TL) to that of a commonly used controller (linear discriminant analysis, or LDA). The participants wore a simulated prosthesis and performed complex functional tasks across multiple limb positions. Analysis resulting from our suite of metrics identified 16 instances of a user-facing problem known as the "limb position effect". We determined that RCNN-TL performed the same as or significantly better than LDA in four such problem instances. We also confirmed that transfer learning can minimize user training burden. Overall, this study contributes a multifaceted new suite of control evaluation metrics, along with a guide to their application, for use in research and testing of myoelectric controllers today, and potentially for use in broader rehabilitation technologies of the future.

Analytic validation of convolutional neural network-based classification of pigmented skin lesions (PSLs) using unseen PSL hyperspectral data for clinical applications

Analytic validation of convolutional neural network-based classification of pigmented skin lesions (PSLs) using unseen PSL hyperspectral data for clinical applications

A prior knowledge-guided distributionally robust optimization-based adversarial training strategy for medical image classification

Medical image classification plays a vital role in computer vision applications within the field of healthcare and medicine, and deep neural network-based classifiers are continuously achieving new breakthroughs and demonstrating tremendous potential in medical imaging analysis. However, the lack of robustness in deep learning techniques makes it risky to apply these classifiers to the domain of healthcare. In addition, the existing adversarial training strategies and domain generation methods are difficult to generalize into the medical imaging field challenged by complex medical texture features. To address this issue, we propose a medical morphological knowledge-guided adversarial training strategy, by jointly considering the robustness against medical data distribution shifts and adversarial attacks from the view of distributionally robust optimization. First, we train a surrogate model with the augmented dataset by guided filtering for capturing model attention on medical morphological information. Next, we design a gradient normalization-based prior knowledge injection module to transfer the attention information learned by surrogate model to the main classifier. Finally, we design a distributionally robust a optimization-based training strategy to induce the main classifier to learn key diagnostic clues as well as enhance the robustness against adversarial attacks. To evaluate the effectiveness of the proposed methods, we perform experiments on two types of in-domain and out-of-domain medical image sets, which contain lung CT scan datasets and dermatoscopic image datasets. Comparative results show that the proposed training strategy achieves higher adversarial attack accuracy than all involved state-of-the-art adversarial training methods and domain generation methods. The code is available at https://github.com/sysu19351146/MMK-DRO.

A Test Statistic Estimation-Based Approach for Establishing Self-Interpretable CNN-Based Binary Classifiers.

Interpretability is highly desired for deep neural network-based classifiers, especially when addressing high-stake decisions in medical imaging. Commonly used post-hoc interpretability methods have the limitation that they can produce plausible but different interpretations of a given model, leading to ambiguity about which one to choose. To address this problem, a novel decision-theory-inspired approach is investigated to establish a self-interpretable model, given a pre-trained deep binary black-box medical image classifier. This approach involves utilizing a self-interpretable encoder-decoder model in conjunction with a single-layer fully connected network with unity weights. The model is trained to estimate the test statistic of the given trained black-box deep binary classifier to maintain a similar accuracy. The decoder output image, referred to as an equivalency map, is an image that represents a transformed version of the to-be-classified image that, when processed by the fixed fully connected layer, produces the same test statistic value as the original classifier. The equivalency map provides a visualization of the transformed image features that directly contribute to the test statistic value and, moreover, permits quantification of their relative contributions. Unlike the traditional post-hoc interpretability methods, the proposed method is self-interpretable, quantitative. Detailed quantitative and qualitative analyses have been performed with three different medical image binary classification tasks.

CR2Net: A Neural Network-Based Classifier for Rician and Rayleigh Channels via Hankelization

CR2Net: A Neural Network-Based Classifier for Rician and Rayleigh Channels via Hankelization

Fingerprint Identification with Fusion of Gabor and Minutiae Features Using BPNN Classifier

This study presents an innovative approach to fingerprint identification by leveraging the fusion of Gabor and Minutiae features, employing a Backpropagation Neural Network (BPNN) classifier for accurate categorization. The process involves initial handling of a fingerprint image dataset, including crucial pre-processing steps such as resizing images and implementing morphological operations like dilation, erosion, and opening. Subsequently, Gabor features and minutiae extraction are performed, followed by the fusion of these features to create a comprehensive representation. To address dimensionality concerns, Principal Component Analysis (PCA) is applied. The dataset, comprising the fused features and corresponding labels, is then loaded for the final step - classification using a BPNN. The network is configured for feed-forward backpropagation, distinguishing fingerprint patterns into categories such as arch, left loop, right loop, tented, and whorl. The evaluation metric used to measure the success of the classification process is accuracy. This approach aims to enhance fingerprint recognition by combining distinctive Gabor and minutiae features, ultimately achieving a more robust and precise identification system through the utilization of neural network-based classification. Keywords—BPNN, PCA, Finger Print Images Dataset

Distinguishing methicillin-resistant Staphylococcus aureus from methicillin-sensitive strains by combining Fe3O4 magnetic nanoparticle-based affinity mass spectrometry with a machine learning strategy

Pathogenic bacteria, including drug-resistant variants such as methicillin-resistant Staphylococcus aureus (MRSA), can cause severe infections in the human body. Early detection of MRSA is essential for clinical diagnosis and proper treatment, considering the distinct therapeutic strategies for methicillin-sensitive S. aureus (MSSA) and MRSA infections. However, the similarities between MRSA and MSSA properties present a challenge in promptly and accurately distinguishing between them. This work introduces an approach to differentiate MRSA from MSSA utilizing matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) in conjunction with a neural network-based classification model. Four distinct strains of S. aureus were utilized, comprising three MSSA strains and one MRSA strain. The classification accuracy of our model ranges from ~ 92 to ~ 97% for each strain. We used deep SHapley Additive exPlanations to reveal the unique feature peaks for each bacterial strain. Furthermore, Fe3O4 MNPs were used as affinity probes for sample enrichment to eliminate the overnight culture and reduce the time in sample preparation. The limit of detection of the MNP-based affinity approach toward S. aureus combined with our machine learning strategy was as low as ~ 8 × 103 CFU mL−1. The feasibility of using the current approach for the identification of S. aureus in juice samples was also demonstrated.Graphical

Convolutional neural network-based classification of glaucoma using optic radiation tissue properties

BackgroundSensory changes due to aging or disease can impact brain tissue. This study aims to investigate the link between glaucoma, a leading cause of blindness, and alterations in brain connections.MethodsWe analyzed diffusion MRI measurements of white matter tissue in a large group, consisting of 905 glaucoma patients (aged 49-80) and 5292 healthy individuals (aged 45-80) from the UK Biobank. Confounds due to group differences were mitigated by matching a sub-sample of controls to glaucoma subjects. We compared classification of glaucoma using convolutional neural networks (CNNs) focusing on the optic radiations, which are the primary visual connection to the cortex, against those analyzing non-visual brain connections. As a control, we evaluated the performance of regularized linear regression models.ResultsWe showed that CNNs using information from the optic radiations exhibited higher accuracy in classifying subjects with glaucoma when contrasted with CNNs relying on information from non-visual brain connections. Regularized linear regression models were also tested, and showed significantly weaker classification performance. Additionally, the CNN was unable to generalize to the classification of age-group or of age-related macular degeneration.ConclusionsOur findings indicate a distinct and potentially non-linear signature of glaucoma in the tissue properties of optic radiations. This study enhances our understanding of how glaucoma affects brain tissue and opens avenues for further research into how diseases that affect sensory input may also affect brain aging.

Optimal Artificial Neural Network-based Fabric Defect Detection and Classification

Automated Fabric Defect (FD) detection plays a crucial role in industrial automation within fabric production. Traditionally, the identification of FDs heavily relies on manual assessment, facilitating prompt repairs of minor defects. However, the efficiency of manual recognition diminishes significantly as labor working hours increase. Consequently, there is a pressing need to introduce an automated analysis method for FD recognition to reduce labor costs, minimize errors, and improve fabric quality. Many researchers have devised defect detection systems utilizing Machine Learning (ML) approaches, enabling swift, accurate, and efficient identification of defects. This study presents the Optimal Artificial Neural Network-based Fabric Defect Detection and Classification (OANN-FDDC) technique. The OANN-FDDC technique exploits handcrafted features with a parameter-tuning strategy for effectively detecting the FD process. To obtain this, the OANN-FDDC technique employs CLAHE and Bilateral Filtering (BF) model-based contrast augmentation and noise removal. Besides, the OANN-FDDC technique extracts shape, texture, and color features. For FD detection, the ANN method is utilized. To improve the detection results of the ANN method, the Root Mean Square Propagation (RMSProp) optimization technique is used for the parameter selection process. The simulation outputs of the OANN-FDDC technique were examined on an open fabric image database. The experimental results of the OANN-FDDC technique implied a better outcome than the 96.97% of other recent approaches.

DIAT-DSCNN-ECA-Net: separable convolutional neural network-based classification of galaxy morphology

DIAT-DSCNN-ECA-Net: separable convolutional neural network-based classification of galaxy morphology

Enhancing Convolutional Neural Networks via separately trained kernels for digit recognition

This paper introduces a novel methodology for pre-training neural networks. Instead of the traditional approach of running a single classifier for all objects, the method used in the research constructs individual classifiers for each object to extract unique features. These classifiers, in essence, act as a series of binary classifiers, each pre-training a distinct set of convolutional kernels. These individually trained kernels are then combined and processed by a comprehensive classifier. The methodology leverages the power of individual feature extraction and collaborative processing to enhance the overall performance of the classifier. The results demonstrate that this innovative pre-training approach leads to a significant improvement in classification performance, yielding higher accuracy compared to conventional methods. It underscores the benefits of incorporating individual object feature extraction and combined processing in the pre-training phase of neural network-based classification tasks. The study opens new paths for improving pre-training methods in neural networks, with potential applications in various fields that require high-accuracy object recognition. Future work will delve deeper into the potential challenges associated with model complexity and overfitting, and will also include a more comprehensive evaluation using extensive training, cross-validation, and independent test sets. This will further validate the effectiveness of our proposed pre-training methodology.

Analysis of Human Voice for Speaker Recognition: Concepts and Advancement

Human voice or speech is a contactless, non-invasive biometric trait for human recognition, easy to use with minimal computer complexity and inexpensive to implement. Speaker recognition (SR) has turned out to be a magnificent approach using speech as the central premise since decades. Its broad range of usages, like forensic speech verification to identify culprits by law enforcement authorities and access control to mobile banking, mobile shopping, etc., has made it a lucrative area of research. Also, the ease of use and dependability of SR will significantly assist people with disabilities in securely accessing and reaping the benefits of digital-era services. Additionally, the emergence of numerous deep learning methods for feature extraction and classification, has helped SR to achieve tremendous progress. This paper presents a comprehensive study on the progression of SR for decades till the present, including integration with Blockchain and challenges. It covers most of the factors that influence SR performance such as fundamentals and structure of SR, different speech pre-processing techniques, various speech features, feature extraction techniques, traditional and neural network-based classification techniques and deep learning-based SR toolkits. As a consequence, in this digital Blockchain era, it will help to design robust and reliable recognition-based services for mankind.

Research on the application of graph neural networks in text classification

With the rapid development of the Internet and social media, a large amount of text data is constantly generated, and text classification has become an important task. Text classification is an important branch in the field of natural language processing and has attracted the attention of many researchers in recent years. Due to the ease of editing text data, most texts in the network are manually constructed and uploaded by users. Therefore, the standardization of online texts and the classification of texts with different granularity are of great significance in the field of information retrieval. This article introduces the basic concepts of text classification and three text classification methods: knowledge engineering (rule) based text classification methods, traditional machine learning-based text classification methods, and deep neural network-based text classification methods. This article also analyzes the advantages and disadvantages of the three methods. Subsequently, the basic concepts and common models of graph neural networks were explored. An application case of graph neural networks in text classification is provided. This article provides effective suggestions for text classification applications in the field of natural language processing.