Classification Results Research Articles

Few‐Shot Contrastive Learning‐Based Multi‐Round Dialogue Intent Classification Method

ABSTRACTTraditional text classification models face challenges in handling long texts and understanding topic transitions in dialogue scenarios, leading to suboptimal performance in automatic speech recognition (ASR)‐based multi‐round dialogue intent classification. In this article, we propose a few‐shot contrastive learning‐based multi‐round dialogue intent classification method. First, the ASR texts are partitioned, and role‐based features are extracted using a Transformer encoder. Second, refined sample pairs are forward‐propagated, adversarial samples are generated by perturbing word embedding matrices and contrastive loss is applied to positive sample pairs. Then, positive sample pairs are input into a multi‐round reasoning module to learn semantic clues from the entire scenario through multiple dialogues, obtain reasoning features, input them into a classifier to obtain classification results, and calculate multi‐task loss. Finally, a prototype update module (PUM) is introduced to rectify the biased prototypes by using gated recurrent unit (GRU) to update the prototypes stored in the memory bank and few‐shot learning (FSL) task. Experimental evaluations demonstrate that the proposed method outperforms state‐of‐the‐art methods on two public datasets (DailyDialog and CM) and a private real‐world dataset.

Implementation of an Image Tampering Detection System with a CMOS Image Sensor PUF and OP-TEE

Since image recognition systems use image data acquired by image sensors for analysis by AI technology, an important security issue is guaranteeing the authenticity of data transmitted from image sensors to successfully perform inference using AI. There have been reports of physical attacks on image sensor interfaces by tampering with images to cause misclassifications in AI classification results. As a countermeasure against these attacks, it is effective to add authenticity to image data with a message authentication code (MAC). For the implementation of this, it is important to have technologies for generating MAC keys on image sensors and to create an environment for secure MAC verification on the host device. For MAC key generation, we used the CIS-PUF technology, which generates MAC keys from PUF responses and random numbers from CMOS image sensor variations. For the secure MAC verification, we used TEE technology, which executes security-critical processes in an environment isolated from the normal operating system. In this study, we propose and demonstrate an image tampering detection system based on MAC verification with CIS-PUF and OP-TEE in an open portable TEE on an ARM processor. In the experiments, we demonstrated a system that computed and transmitted MAC for captured images using the CIS-PUF key and then performed MAC verification in the secure world of the OP-TEE.

TATPat based explainable EEG model for neonatal seizure detection

The most cost-effective data collection method is electroencephalography (EEG) to obtain meaningful information about the brain. Therefore, EEG signal processing is very important for neuroscience and machine learning (ML). The primary objective of this research is to detect neonatal seizures and explain these seizures using the new version of Directed Lobish. This research uses a publicly available neonatal EEG signal dataset to get comparative results. In order to classify these EEG signals, an explainable feature engineering (EFE) model has been proposed. In this EFE model, there are four essential phases and these phases: (i) automaton and transformer-based feature extraction, (ii) feature selection deploying cumulative weight-based neighborhood component analysis (CWNCA), (iii) the Directed Lobish (DLob) and Causal Connectome Theory (CCT)-based explainable result generation and (iv) classification deploying t algorithm-based support vector machine (tSVM). In the first phase, we have used a channel transformer to get channel numbers and these values have been divided into three levels and these levels are named (1) high, (2) medium and (3) low. By utilizing these levels, we have created an automaton and this automaton has three nodes (each node defines each level). In the feature extraction phase, transition tables of these nodes has been extracted. Therefore, the proposed feature extraction function is termed Triple Nodes Automaton-based Transition table Pattern (TATPat). The used EEG signal dataset contains 19 channels and there are 9 (= 32) connection in the defined automaton. Thus, the presented TATPat extracts 3249 (= 19 × 19 × 9) features from each EEG segment. To choose the most informative features of these 3249 features, a new feature selector which is CWNCA has been applied. By cooperating findings of this feature selector and the presented DLob, the explainable results have been obtained. The last phase is the classification phase and to get high classification performance from this phase, an ensemble classifier (tSVM) has been presented and the classification results have been obtained using two validation techniques which are 10-fold cross-validation (CV) and leave-one subject-out (LOSO) CV. The proposed EFE model generates a DLob string and by using this string, the explainable results have been obtained. Moreover, the presented EFE model attained 99.15% and 76.37% classification accuracy deploying 10-fold and LOSO CVs respectively. According to the classification performances, the recommended TATPat-based EFE is a good model at EEG signal classification. Also, the presented TATPat-based EFE model is a good model for explainable artificial intelligence (XAI) since TTPat-based EFE is cooperating by the DLob.

High frequency radar error classification and prediction based on K-means methods

This study aims to characterize the high frequency radar and numerically simulated low-frequency filtered currents in the south-eastern Bay of Biscay (study area) using a K-means classification algorithm based on an improved Euclidean Distance calculation method that does not take missing values. The errors between observations and simulations was estimated and predicted based on this classification method. Results indicate that predominantly eastward (northward) currents over the Spanish (French) continental shelf/slope in winter and more variable currents in the west and south-west in summer. The model classification results for circulation characteristics are in relatively good agreement with HF radar results, especially for currents on the Spanish (French) shelf/slope. In addition, the probabilistic relationship between observed and modeled currents was explored, obtaining the probability of occurrence of modeled current groups when each group of observed currents occurs. Finally, predictions of model and observed current errors were made based on the classification results, and it was found that the predictions based on the classification of all data had the smallest errors, with a 17% improvement over the unclassified control experiment. This study provides a foundation for subsequent model error testing, forecast product improvement and data assimilation.

Motor Imagery Classification Improvement of Two-Class Data with Covariance Decentering Eigenface Analysis for Brain–Computer Interface Systems

This study is intended to improve the motor imagery classification performance of two-class data points using newly developed covariance decentering eigenface analysis (CDC-EFA). When extracting the classification for the given data points, it is necessary to precisely distinguish the classes because the left and right features are difficult to differentiate. However, when centering is performed, the unique average data of each feature are lost, making them difficult to distinguish. CDC-EFA reverses the centering method to enhance data characteristics, making it possible to assign weights to data with a high correlation with other data. In experiments with the BCI dataset, the proposed CDC-EFA method was used after preprocessing by filtering and selecting the electroencephalogram data. The decentering process was then performed on the covariance matrix calculated when acquiring the unique face. Subsequently, we verified the classification improvement performance via simulations using several BCI competition datasets. Several signal processing methods were applied to compare the accuracy results of the motor imagery classification. The proposed CDC-EFA method yielded an average accuracy result of 98.89%. Thus, it showed improved accuracy compared with the other methods and stable performance with a low standard deviation.

Deep Learning-Based Pulmonary Nodule Screening: A Narrative Review

AbstractGiven its capacity to generate three-dimensional pictures, computed tomography is the most effective means of detecting lung nodules with more excellent resolution of detected nodules. Small lung nodules can easily be overlooked on chest X-rays, making interpretation difficult. Artificial intelligence algorithms have recently demonstrated remarkable progress in medical imaging, especially with deep learning techniques such as convolutional neural networks (CNNs). CNN produces excellent results in natural image recognition and classification using abundant available data and the computational abilities of modern computers. It further reduces false-positive pulmonary nodules in medical image processing. This review article provides a detailed and inclusive review of recent advances, challenges, performance comparisons, and possible future directions for the problem of pulmonary nodule screening using deep learning methods.

On quaternionic bisectional curvature

AbstractIn this article we study the concept of quaternionic bisectional curvature introduced by Chow and Yang in (J Differ Geom 29(2):361–372, 1989) for quaternion-Kähler manifolds. We show that non-negative quaternionic bisectional curvature is only realized for the quaternionic projective space $$\mathbb {H}\textrm{P}^n$$ H P n . We also show that all symmetric quaternion-Kähler manifolds different from $$\mathbb {H}\textrm{P}^n$$ H P n admit quaternionic lines of negative quaternionic bisectional curvature. In particular this implies that non-negative sectional curvature does not imply non-negative quaternionic bisectional curvature. Moreover we give a new and rather short proof of a classification result by A. Gray on compact Kähler manifolds of non-negative sectional curvature.

Segmentation-based truncated-SVD for effective feature extraction in hyperspectral image classification

ABSTRACT Remote sensing hyperspectral images (HSIs) are rich sources of information about land cover captured across hundreds of narrow, contiguous spectral wavelength bands. However, using the entire original HSI for practical applications can lead to suboptimal classification accuracy. To address this, band reduction techniques, categorized as feature extraction and feature selection methods, are employed to enhance classification results. One commonly used feature extraction approach for HSIs is Principal Component Analysis (PCA). However, PCA may fall short of capturing the local and specific characteristics present in the HSI data. In this paper, we introduce two novel feature extraction methods: Segmented Truncated Singular Value Decomposition (STSVD) and Spectrally Segmented Truncated Singular Value Decomposition (SSTSVD) to improve classification performance. Segmentation is carried out based on highly correlated bands’ segments and spectral bands’ segments within the HSI data. Our study evaluates and compares these newly proposed methods against classical feature extraction methods, including PCA, Incremental PCA, Sparse-PCA, Kernel PCA, Segmented-PCA (SPCA), and Truncated Singular Value Decomposition (TSVD). We perform this analysis on three distinct HSI datasets, namely the Indian Pines HSI, the Pavia University HSI, and the Kennedy Space Center HSI, using per-pixel Support Vector Machine (SVM) and Random Forest (RF) classification. The experimental results demonstrate the superiority of our proposed methods for all three datasets. The best-performing feature extraction methods when classification is performed using an SVM classifier are STSVD3 (89.03%), SSTSVD2 (95.55%), and STSVD3 (97.74%) for the Indian Pines, Pavia University, and Kennedy Space Center datasets, respectively. Similarly, for the RF classifier, the best-performing feature extraction methods are SSTSVD4 (88.98%), SSTSVD3 (96.04%), and SSTSVD4 (96.09%) for Indian Pines, Pavia University, and Kennedy Space Center datasets, respectively.

Forensic bone age assessment of hand and wrist joint MRI images in Chinese han male adolescents based on deep convolutional neural networks.

In Chinese criminal law, the ages of 12, 14, 16, and 18 years old play a significant role in the determination of criminal responsibility. In this study, we developed an epiphyseal grading system based on magnetic resonance image (MRI) of the hand and wrist for the Chinese Han population and explored the feasibility of employing deep learning techniques for bone age assessment based on MRI of the hand and wrist. This study selected 282 Chinese Han Chinese males aged 6.0-21.0 years old. In the course of our study, we proposed a novel deep learning model for extracting and enhancing MRI hand and wrist bone features to enhance the prediction of target MRI hand and wrist bone age and achieve precise classification of the target MRI and regression of bone age. The evaluation metric for the classification model including precision, specificity, sensitivity, and accuracy, while the evaluation metrics chosen for the regression model are MAE. The epiphyseal grading was used as a supervised method, which effectively solved the problem of unbalanced sample distribution, and the two experts showed strong consistency in the epiphyseal plate grading process. In the classification results, the accuracy in distinguishing between adults and minors was 91.1%, and the lowest accuracy in the three minor classifications (12, 14, and 16 years of age) was 94.6%, 91.1% and 96.4%, respectively. The MAE of the regression results was 1.24 years. In conclusion, the deep learning model proposed enabled the age assessment of hand and wrist bones based on MRI.

Lumbar Spinal Stenosis Grading in Multiple Level Magnetic Resonance Imaging Using Deep Convolutional Neural Networks.

Retrospective magnetic resonance imaging grading with comparison between experts and deep convolutional neural networks (CNNs). The application of deep learning to clinical diagnosis has gained popularity. This approach can accelerate image interpretation and serve as a screening tool to help doctors. A comparison was conducted between retrospective magnetic resonance imaging (MRI) grading performed by experts and grading obtained using CNN classifiers. Data were collected from the lumbar axial dataset in the DICOM format. Two experts labeled the sampled images using the same diagnostic tools: localization of patches near the spinal canal, rootlet leveling, and stenosis grading. Comprehensive comparisons were presented for both rootlet cord classification and stenosis grading. Rootlet-cord classification for the two analyzers was 90.3% and the F1 score was 86.6%. The agreement of Analyzers-Classifiers was 92.7% and 96.8% for data with 90.6% and 95.6% F1 scores, respectively. For stenosis grading, there was an agreement of 89.2% between the two analyzers, resulting in an F1 score of 76.5%. The grades of the Analyzers-Classifiers agreed on 91.5/89.4% of the data, with an F1 score of 78.4/75.7%. Analyzer1 and Analyzer2 classified >74% as grade A (78.8% and 74.4%, respectively), 15.4% and 18.6% as grade B, 4.2% and 6.0% as grade C, and 1.6% and 2.0% as grade D, respectively. The fully automated deep learning model showed competitive results in stenosis grade diagnosis and rootlet cord classification under similar anatomical conditions. However, abrupt anatomical changes can lead to a puzzle diagnosis based only on images.

Assessing Data Fusion in Sensory Devices for Enhanced Prostate Cancer Detection Accuracy

The combination of an electronic nose and an electronic tongue represents a significant advance in the pursuit of effective detection methods for prostate cancer, a widespread form of cancer affecting men across the globe. These cutting-edge devices, collectively called “E-Senses”, use data fusion to identify distinct chemical compounds in exhaled breath and urine samples, potentially improving existing diagnostic techniques. This study combined the information from two sensory perception devices to detect prostate cancer in biological samples (breath and urine). To achieve this, data from patients diagnosed with the disease and from control individuals were collected using a gas sensor array and chemical electrodes. The signals were subjected to data preprocessing algorithms to prepare them for analysis. Following this, the datasets for each device were individually analyzed and subsequently merged to enhance the classification results. The data fusion was assessed and it successfully improved the accuracy of detecting prostate-related conditions and distinguishing healthy patients, achieving the highest success rate possible (100%) in classification through machine learning methods, outperforming the results obtained from individual electronic devices.

Riding Feeling Recognition Based on Multi-Head Self-Attention LSTM for Driverless Automobile

With the emergence of driverless technology, passenger ride comfort has become an issue of concern. In recent years, driving fatigue detection and braking sensation evaluation based on EEG signals have received more attention, and analyzing ride comfort using EEG signals is also a more intuitive method. However, it is still a challenge to find an effective method or model to evaluate passenger comfort. In this paper, we propose a long- and short-term memory network model based on a multiple self-attention mechanism for passenger comfort detection. By applying the multiple attention mechanism to the feature extraction process, more efficient classification results are obtained. The results show that the long- and short-term memory network using the multi-head self-attention mechanism is efficient in decision making along with higher classification accuracy. In conclusion, the classifier based on the multi-head attention mechanism proposed in this paper has excellent performance in EEG classification of different emotional states, and has a broad development prospect in brain-computer interaction.

Identification of asymmetrical abnormalities in functional connectivity and brain network topology for migraine sufferers: A preliminary study based on resting-state fMRI data

Research on the neural mechanisms underlying brain asymmetry in patients with migraine patients using fMRI is insufficient. This study proposed using lateralized algorithms for functional connectivity and brain network topology and investigated changes in their lateralization in patients with migraine. In study 1, laterality indices of functional connectivity (LFunctionCorr) and brain network topological properties (LBetweennessCentrality, LDegree, and LStrength) were defined. Differences between migraineurs and normal subjects were compared at whole-brain, half-brain, and region levels. In study 2, laterality indices were used to classify migraine and were validated using independent samples and the segment method for repeatability. In study 3, abnormal brain regions related to migraine were extracted based on the classification results and differences analysis. Study 1 found no significant differences related to in for migraine at the whole-brain level; however, significant differences were identified at the half-brain level for the hemispheric lateralization of the LFunctionCorr, while 11 significantly different brain regions were also identified at the brain region level. Furthermore, the classification accuracy in study 2 was 0.9366. With repeated validation, the accuracy reached 0.8561. Furthermore, after extending the samples according to the segmentation strategy, the classification accuracies were improved to 0.9408 and 0.8585. Study 3 identified 10 crucial brain regions with asymmetrical specificity based on laterality indices distributed across the visual network, the frontoparietal control network, the default mode network, the salience/ventral attention network and the limbic system. The results revealed novel insights and avenues for research into the mechanisms of migraine asymmetry and showed that the laterality indices could be used as a potential diagnostic imaging marker for migraine.

Detection of clavicle fracture after martial arts training based on 3D segmentation and multi-perspective fusion

To improve the diagnosis efficiency of clavicle fractures, a Diagnosis Model of Clavicle Fracture Based on 3D Segmentation and Multi Perspective Fusion (3Ds MPF) is proposed in the experiment. By performing 3D segmentation on clavicle images, key layer images are extracted, and multi-perspective image data are fused and output to achieve accurate classification and diagnosis results. The results show that on the relevant dataset, the loss function value of the constructed model is less than 0.01, and the calculation accuracy is as high as 0.999. The average classification accuracy in actual use exceeds 90%, and it can capture the majority of fracture features in clavicle computed tomography images. The above results indicate that the model constructed by the research institute can significantly improve the judgment efficiency of computed tomography images, which helps to enhance the detection and diagnosis of clinical fracture conditions. The experimental results have helped improve the clinical diagnostic efficiency of fracture detection work.

Center-Highlighted Multiscale CNN for Classification of Hyperspectral Images

Hyperspectral images (HSIs) capture a wide range of spectral features across multiple bands of light, from visible to near-infrared. Hyperspectral image classification technology enables researchers to accurately identify and analyze the composition and distribution of surface materials. Current mainstream deep learning methods typically use block sampling to capture spatial features for the model. However, this approach can affect classification results due to the influence of neighboring features within the sample block. To improve the model’s focus on the center of the sampling block, this study proposes a center highlight with multiscale CNN for hyperspectral image classification (CHMSC). The network utilizes an automatic channel selector (Auto-CHS) to fully consider every channel feature and capture the correlation between the channels. Then, CHMSC enhances the model’s ability to concentrate on the central features of the sampling block utilizing structures such as the center highlight. Finally, before outputting the prediction results, an SENet is employed to further refine the features and learn associate interactions between different scales of spatial features and spectral features. Experimental results from three hyperspectral datasets validate the effectiveness of the proposed method. Specifically, when 15 samples from each class are selected for training, CHMSC achieves the highest overall accuracy (OA) of 90.05%, 92.78%, and 90.15% on the three datasets, outperforming other methods with increases of more than 3.11%, 1.8%, and 2.01% in OA, respectively.

Using a Yolov8-Based Object Detection Model for an Automatic Garbage Sorting System

This study presents the development of an automatic garbage sorting system based on computer vision techniques and the You Only Look Once version 8 (YOLOv8) algorithm. RGB images of garbage, including cans and plastic bottles, were collected by a camera. The YOLOv8 model, trained on a dataset of both RGB images of cans and plastic bottles, acts as the core for detection and classification. During real-time sorting, the classification results and picking points from the YOLOv8 model are used by a gantry robot. The gantry robot is developed and controlled by a Programmable Logic Controller (PLC) to place each type of garbage into its respective categories. A conveyor tracking algorithm was implemented to pick up garbage moving on the conveyor. The system's performance is evaluated using 20 garbage samples for the can and the plastic bottle category. The accuracy, precision, recall, and F1 score were 92.5%, 90.5%, 95%, and 92.7%, respectively. The system proposed in this study can be modified to sort other garbage types.

Automatic plant phenotyping analysis of Melon (Cucumis melo L.) germplasm resources using deep learning methods and computer vision

Cucumis melo L., commonly known as melon, is a crucial horticultural crop. The selection and breeding of superior melon germplasm resources play a pivotal role in enhancing its marketability. However, current methods for melon appearance phenotypic analysis rely primarily on expert judgment and intricate manual measurements, which are not only inefficient but also costly. Therefore, to expedite the breeding process of melon, we analyzed the images of 117 melon varieties from two annual years utilizing artificial intelligence (AI) technology. By integrating the semantic segmentation model Dual Attention Network (DANet), the object detection model RTMDet, the keypoint detection model RTMPose, and the Mobile-Friendly Segment Anything Model (MobileSAM), a deep learning algorithm framework was constructed, capable of efficiently and accurately segmenting melon fruit and pedicel. On this basis, a series of feature extraction algorithms were designed, successfully obtaining 11 phenotypic traits of melon. Linear fitting verification results of selected traits demonstrated a high correlation between the algorithm-predicted values and manually measured true values, thereby validating the feasibility and accuracy of the algorithm. Moreover, cluster analysis using all traits revealed a high consistency between the classification results and genotypes. Finally, a user-friendly software was developed to achieve rapid and automatic acquisition of melon phenotypes, providing an efficient and robust tool for melon breeding, as well as facilitating in-depth research into the correlation between melon genotypes and phenotypes.

Quantitative characterisation of debris cloud-induced pitting damage in AL-Whipple shields of spacecraft using infrared thermography

ABSTRACT Quantitatively evaluating the debris cloud-induced pitting damage in the AL-Whipple structure of spacecraft, is still a challenge, let alone microstructural plastic deformation usually accompanied by geometric cratering defects, which may decrease material stress-carrying capacity and structural life. With this motivation, a dedicated modelling technique is proposed to gain an insight into various modalities (i.e., geometric and plastic defects) of pitting damage-induced thermal modulation for infrared inspection. Then, a sequence of infrared images is obtained and analysed using the gap smoothing algorithm and naive Bayes classification methods. On this basis, a quantitative characterization framework is proposed to characterize the damage modalities and their morphologies. This method is validated using a 3D numerical model which contains pitting damage acquired from the real-world structure. Modality classification and morphology evaluation results are well consistent with the artificial defects, in terms of the modality, diameter, and depth. Then they‘re experimentally corroborated, in which the severity and distribution of pitting damage are precisely characterized, in terms of the crater number and morphologies, as well as the size of plastic regions around these craters, which are consistent with the actual observations. This work proposes a non-destructive strategy for pitting damage evaluation, which is engineering-friendly.

A Hierarchical Coarse–Fine Adaptive Fusion Network for the Joint Classification of Hyperspectral and LiDAR Data

Hyperspectral image (HSI) and light detection and ranging (LiDAR) data have gained significant attention due to their excellent complementarity, which can be combined to improve the accuracy of land cover classification. However, there are still many challenges, such as unrelated physical characteristics, different data structures, and a lack of labeled samples. Many methods fail to leverage the full potential of multi-source data, particularly hierarchical complementary information. To address these problems, a hierarchical coarse–fine adaptive (HCFA) fusion network with dynamic convolution and a transformer is proposed for multi-source remote sensing land cover classification. Although fusing hierarchical information can improve the classification accuracy of the model, improper hierarchical feature selection and optimization may negatively affect the classification results. Therefore, a coarse–fine mutual learning strategy is proposed to dynamically fuse hierarchical information. Additionally, the disparity between multi-source data continues to prevent the realization of effective fusion. To tackle this challenge, cross-tokenization and cross-token attention are implemented to enhance information interaction. Furthermore, to improve the model representation with limited computational cost, we combine the advantages of dynamic convolution with a transformer. Validation on three standard datasets demonstrates that HCFA achieves high accuracy with just 1% of the training set while maintaining low computational costs.

The Adversarial Attacks towards Human Action Recognition Models: A Comparison between unimodal and multimodal Models

Human-action recognition models are neural networks that analyse visual inputs and provide classification or text outputs. This technology has and will significantly impact society in security, education, healthcare, etc. However, Human-Action Recognition models, like other neural networks, are still susceptible to malicious adversarial attacks. Therefore, this paper proposes an experimental adversarial attack towards ResNet-18 using FGSM. First, ResNet-18 is finetuned using the UCF-101 dataset, and keyframes are selected from sample videos. The keyframes will be given to ResNet-18 for classification while FGSM will be implemented, and ResNet will do another classification of the attached sample. The classification results (Original and Attacked) are given to the language model (GPT-4o) through a prompt that provides the language model with a specific role (e.g. a smart home assistant), and this section is regarded as unimodal. The original and attacked frames will be sent directly instead of the labels in the multimodal section. Lastly, this paper proposes to observe the effects on textual responses generated based on a given prompt and the classification result and evaluate the impact of the attack through Cosine Similarities and Human Evaluation.