Test Accuracy Research Articles

A new deep learning model with interface for fine needle aspiration cytology image-based breast cancer detection

Cytological evaluation through microscopic image analysis of fine needle aspiration cytology (FNAC) is pivotal in the initial screening of breast cancer. The sensitivity of FNAC as a screening tool relies on both image quality and the pathologist’s expertise. To enhance diagnostic accuracy and alleviate the pathologist’s workload, a computer-aided diagnosis (CAD) system was developed. A comparative study was conducted, assessing twelve candidate pre-trained models. Utilizing a locally gathered FNAC image dataset, three superior models-MobileNet-V2, DenseNet-121, and Inception-V3-were selected based on their training, validation, and testing accuracies. Further, these models underwent evaluation in four transfer learning scenarios to enhance testing accuracy. While the outcomes were promising, they left room for improvement, motivating us to create a novel deep convolutional neural network (CNN). The newly proposed model exhibited robust performance with testing accuracy at 85%. Our research concludes that the most lightweight, high-accuracy model is the one we propose. We’ve integrated it into our user-friendly Android App, “Breast Cancer Detection System,” in TensorFlow Lite format, with cloud database support, showcasing its effectiveness. Implementing an artificial intelligent (AI)-based diagnosis system with a user-friendly interface holds the potential to enhance early breast cancer detection using FNAC.

Comparison of WASPAS and VIKOR methods to determine non-cash food assistance recipients

Non-cash food assistance or bantuan pangan non-tunai (BPNT) is a government program of the Republic of Indonesia by distributes food assistance in non-cash to beneficiary families. The process of distributing BPNT still needs to be done with the data and criteria set, because the existing BPNT distribution is considered not right on target. We need a method that can help provide an objective decision. One method that can be used in making decisions is the weighted aggregated sum product assessment (WASPAS) and Vlsekriterijumsko Koompromisno Rangiranje (VIKOR) methods. The results of the calculations from the two methods will then be chosen which is the best, by conducting sensitivity tests and accuracy tests. This study uses 100 sample data and 16 criteria. The sensitivity test results are 9.780678997% for the WASPAS method and -0.0759182% for the VIKOR method, while the results of the accuracy test show that both methods have the same level of accuracy, which is 80%. Based on the comparison of the sensitivity test and accuracy test of the two methods, the WASPAS method is considered more accurate in determining the recipients of the BPNT program because the WASPAS method has a higher sensitivity test value than the VIKOR method.

Handheld echocardiography for the screening and diagnosis of rheumatic heart disease: a systematic review to inform WHO guidelines

Early detection and diagnosis of acute rheumatic fever and rheumatic heart disease are key to preventing progression, and echocardiography has an important diagnostic role. Standard echocardiography might not be feasible in high-prevalence regions due to its high cost, complexity, and time requirement. Handheld echocardiography might be an easy-to-use, low-cost alternative, but its performance in screening for and diagnosing acute rheumatic fever and rheumatic heart disease needs further investigation. In this systematic review and meta-analysis, we searched Embase, MEDLINE, LILACS, and Conference Proceedings Citation Index-Science up to Feb 9, 2024, for studies on the screening and diagnosis of acute rheumatic fever and rheumatic heart disease using handheld echocardiography (index test) or standard echocardiography or auscultation (reference tests) in high-prevalence areas. We included all studies with useable data in which the diagnostic performance of the index test was assessed against a reference test. Data on test accuracy in diagnosing rheumatic heart disease, acute rheumatic fever, or carditis with acute rheumatic fever (primary outcomes) were extracted from published articles or calculated, with authors contacted as necessary. Quality of evidence was appraised using GRADE and QUADAS-2 criteria. We summarised diagnostic accuracy statistics (including sensitivity and specificity) and estimated 95% CIs using a bivariate random-effects model (or univariate random-effects models for analyses including three or fewer studies). Area under the curve (AUC) was calculated from summary receiver operating characteristic curves. Heterogeneity was assessed by visual inspection of plots. This study was registered with PROSPERO (CRD42022344081). Out of 4868 records we identified 11 studies, and two additional reports, comprising 15 578 unique participants. Pooled data showed that handheld echocardiography had high sensitivity (0·87 [95% CI 0·76-0·93]), specificity (0·98 [0·71-1·00]), and overall high accuracy (AUC 0·94 [0·84-1·00]) for diagnosing rheumatic heart disease when compared with standard echocardiography (two studies; moderate certainty of evidence), with better performance for diagnosing definite compared with borderline rheumatic heart disease. High sensitivity (0·79 [0·73-0·84]), specificity (0·85 [0·80-0·89]), and overall accuracy (AUC 0·90 [0·85-0·94]) for screening rheumatic heart disease was observed when pooling data of handheld echocardiography versus standard echocardiography (seven studies; high certainty of evidence). Most studies had a low risk of bias overall. Some heterogeneity was observed for sensitivity and specificity across studies, possibly driven by differences in the prevalence and severity of rheumatic heart disease, and level of training or expertise of non-expert operators. Handheld echocardiography has a high accuracy and diagnostic performance when compared with standard echocardiography for diagnosing and screening of rheumatic heart disease in high-prevalence areas. World Health Organization. For the Chinese, French, Italian, Persian, Portuguese, Spanish and Urdu translations of the abstract see Supplementary Materials section.

Deep learning and quantization for accurate and efficient multi-target radar inference of moving targets

Real-time, radar-based human activity and target classification is useful for wide-area ground surveillance. However, the feasibility of deploying deep learning (DL) models in radar-based systems with limited computational resources remains unexplored. This paper investigated the effect of quantization on model throughput and accuracy for deployment in radar systems. A seven-layer residual network was proposed to classify ground-moving targets and achieved a test accuracy of 87.72%. The model was then quantized to 16-bit and 8-bit precision, resulting in a 3.8 times speedup in inference throughput, with less than a 0.4% drop in test and validation accuracy. The results showed that quantization can improve inference throughput with a negligible decrease in target classification accuracy. The increase in throughput and reduction in computational expense that comes with quantization promotes the feasibility of the deployment of DL models in systems with limited computational resources. The findings of this paper hold significant promise for the successful use of quantized models in modern radar systems, while adhering to stringent size, weight and power consumption constraints.

Enhanced human activity recognition through deep multi-layer perceptron on the UCI-HAR dataset

Using the UCI-HAR dataset, this paper examines human activity recognition (HAR) from the perspectives of data science and artificial intelligence. The primary objective is to present and evaluate the effectiveness of a multi-layer perceptron (MLP) model, concentrating on six different activity categories. We train and assess the MLP model using the UCI-HAR dataset, contrasting its results with those of convolutional neural networks (CNN). The MLP model shows competitive results, attaining an amazing 97% validation and testing accuracy, highlighting its efficiency for smaller datasets. An extensive study is carried out to assess the model's adaptation to a larger Motion Sense dataset using confusion matrices and cross-entropy, the model shows robustness with an accuracy of 89%. The MLP model performs admirably, demonstrating its capacity to pick up complex patterns. Results from comparative analysis with CNN are competitive, especially when dealing with smaller datasets. The suggested MLP model shows up as a practical and efficient way to advance HAR techniques. Its remarkable performance and versatility not only show its usefulness in real-world scenarios but also point to interesting directions for further study in the area of HAR.

Smart measurement and monitoring system for aquaculture fisheries with IoT-based telemetry system

The instrumentation design of an online monitoring device for aquaculture media is discussed in this article. The main processor in this internet of things (IoT) real-time telemetry system is an ESP32 board. Temperature, acidity level, conductivity level, dissolved oxygen (DO) level, and degree of oxygen reduction in the water were the aquaculture parameters measured. The ESP32 collects data from each sensor, groups it into a dataset, displays it on the LCD, saves it to the SD card, and then uploads it to the real-time database. In addition, an Android application is being developed for users. This device has been tested to ensure that each measured parameter is accurate and precise. The accuracy test, one of the major results of laboratory scale tests, demonstrates that each parameter has a different measurement error that represents with average error absolute. Six tested sensors/instruments were subjected to the test. Average absolute error for temperature sensor is +0.76%, pH sensor is +1.52%, electrical conductivity (EC) sensor is +10.8%, oxidation reduction potential (ORP) sensor is +14.6%, DO sensor is +9.3%, and total dissolve solids (TDS) sensor is +13.2%. This device is very dependable and convenient for monitoring the condition of aquaculture media in real-time and accurately.

Use of the Head Impulse, Nystagmus, Test of Skew ('HINTS') assessment to aid differential diagnosis in acute vestibular syndrome in the hyperacute stroke setting.

Differential diagnosis of acute vertigo syndrome is challenging given the similarities between clinical presentations of posterior circulation stroke and peripheral vestibular dysfunction. The Head Impulse, Nystagmus, Test of Skew ('HINTS') assessment is a clinical bedside test used to aid diagnosis. Comprehensive training on use of the Head Impulse, Nystagmus, Test of Skew assessment was provided to one stroke consultant, and the effectiveness of the test in that setting was assessed. Further education was completed with more members of the stroke and emergency department multi-disciplinary team. Quality improvement measures including magnetic resonance imaging use and bed utilisation were explored. Following training of one stroke consultant, the Head Impulse, Nystagmus, Test of Skew assessment was found to be a feasible, accurate bedside test within this acute stroke service. Further training for the multi-disciplinary team was completed, but outcome measures were not explored because of the coronavirus disease 2019 pandemic and maternity leave. There is a role for trained members of the multi-disciplinary team to successfully use the Head Impulse, Nystagmus, Test of Skew assessment in hyperacute stroke settings, to aid diagnosis in acute vertigo syndrome.

Artificial neural networks analysis predicts long-term fistula function in hemodialysis patients following percutaneous transluminal angioplasty

Kidney failure is particularly common in the United States, where it affects over 700,000 individuals. It is typically treated through repeated sessions of hemodialysis to filter and clean the blood. Hemodialysis requires vascular access, in about 70% of cases through an arteriovenous fistula (AVF) surgically created by connecting an artery and vein. AVF take 6 weeks or more to mature. Mature fistulae often require intervention, most often percutaneous transluminal angioplasty (PTA), also known as fistulaplasty, to maintain the patency of the fistula. PTA is also the first-line intervention to restore blood flow and prolong the use of an AVF, and many patients undergo the procedure multiple times. Although PTA is important for AVF maturation and maintenance, research into predictive models of AVF function following PTA has been limited. Therefore, in this paper we hypothesize that based on patient-specific information collected during PTA, a predictive model can be created to help improve treatment planning. We test a set of rich, multimodal data from 28 patients that includes medical history, AVF blood flow, and interventional angiographic imaging (specifically excluding any post-PTA measurements) and build deep hybrid neural networks. A hybrid model combining a 3D convolutional neural network with a multi-layer perceptron to classify AVF was established. We found using this model that we were able to identify the association between different factors and evaluate whether the PTA procedure can maintain primary patency for more than 3 months. The testing accuracy achieved was 0.75 with a weighted F1-score of 0.75, and AUROC of 0.75. These results indicate that evaluating multimodal clinical data using artificial neural networks can predict the outcome of PTA. These initial findings suggest that the hybrid model combining clinical data, imaging and hemodynamic analysis can be useful to treatment planning for hemodialysis. Further study based on a large cohort is needed to refine the accuracy and model efficiency.

Distributed denial of service attacks classification system using features selection and ensemble techniques

Distributed denial-of-service (DDoS) attacks are expanding threat to online services and websites. These attacks overwhelm targets with traffic from multiple sources to exhaust resources and make services unavailable. The frequency of DDoS attacks exhibits an ongoing upward trajectory over time. This persistent escalation highlights the need for effective countermeasures. While machine learning approaches have been extensively investigated for binary classification of DDoS attacks, multi-class classification has received comparatively less examination in the literature despite its greater practical utility. In this paper, we propose an intrusion detection system for detecting and classifying DDoS attacks, based on two main axes: feature selection for selecting the best relevant features and ensemble learning technique for improving performance by combining weak learners. The proposed model has been trained and evaluated on the CICDDoS2019 dataset. Experimental evaluation demonstrates improved performance using a subset of 16 relevant features identified, with a test accuracy of 82.35% attained for discriminating between the 12 classes represented in the dataset. By aggregating attacks sharing common characteristics resulting in 7 classes, the approach achieves surpassing 97% accuracy. Additionally, a binary classification delineating benign and DDoS attacks attain 99.90% accuracy.

Sawtooth-Based FDS for Rapid and Accurate Testing of Transformer Oil–Paper Insulation

Sawtooth-Based FDS for Rapid and Accurate Testing of Transformer Oil–Paper Insulation

Computer model for detecting tsunami wave hazard on built-up land using machine learning and sentinel 2A satellite imagery

<p>The aim of this research is to compile a tsunami wave hazard scale based on built-up land density extracted and classified by machine learning from Sentinel 2A satellite and digital elevation model (DEM) imageries. This research was carried out in 5 stages, namely: (i) pre-processing of Sentinel 2A and DEM images, (ii) Classification of VI data using the machine learning algorithms, (iii) Spatial prediction using the ordinary kriging method, (iv) Field testing using the confusion matrix method, (v) Preparation of decision matrix for tsunami wave hazard. The results of the study show that the most accurate classification algorithm for classifying built-up indices data is the k-nearest neighbor (k-NN) algorithm. The results of the statistical accuracy test show that the most accurate is normalized difference built-up index (NDBI) with a mean of square error (MSE) value of 0.073 and a mean of absolute error (MAE) of 0.003. DEM analysis shows that the research area is at an altitude of 0–15 meters above sea level so it is in the high vulnerability to medium vulnerability category. Field testing showed user accuracy of 91.11%, manufacturer accuracy of 92.16%, and overall average accuracy of 91%.</p>

Traffic Sign Recognition using HIICNN Stack Ensemble Method

Abstract: This project introduces a novel deep convolutional neural network (CNN) model aimed at enhancing traffic sign recognition for autonomous vehicle technology and road safety. Employing a stacking ensemble approach, we combine multiple CNNs with diverse improvement techniques, including preprocessing for adverse weather conditions and shooting errors, and data augmentation to handle class imbalance. Through adjustments in learning rates during model training, we mitigate overfitting. Our ensemble model achieves an exceptional 99.4 percent test accuracy on the German Traffic Sign Recognition Benchmark (GTSRB) dataset, surpassing prior studies. We utilize gradient-weighted Class Activation Mapping (Grad-CAM) for model explain- ability and an evidential deep learning approach to quantify classification uncertainty. Our framework underscores the efficacy of combining preprocessing, ensemble learning, and transfer learning for superior perfor- mance and reliability in traffic sign recognition systems, contributing significantly to autonomous driving and road safety.

Deep ensemble learning for chickenpox detection from clinical images

Chickenpox caused by the varicella-zoster virus is an extremely contagious viral infection common in children and quickly develops into a severe problem. Over 90% of unvaccinated people have been infected. Still, infection occurs at different ages in different parts of the world- Over 70 % of people become infected by the age of 10 years in the United States, the United Kingdom, and Japan, and by the age of 20 in India, West Indies, and South East Asia. Automatic classification of the specific disease is a challenging task to present clinicians to distinguish between different kinds of skin conditions and recommend suitable treatment. Convolutional Neural Networks have recently achieved great success in many machines learning purposes and have presented a state-of-the-art performance in various computer-assisted diagnosis applications. This study proposes a deep neural network-based method that follows an ensemble approach by combining VGG-16, VGG-19, and ResNet-50 architectures to distinguish chickenpox from other skin conditions. Experimental test results have achieved accurate classification with an assuring test accuracy up to 93%.

Evaluating Cellularity Estimation Methods: Comparing AI Counting with Pathologists’ Visual Estimates

The development of next-generation sequencing (NGS) has enabled the discovery of cancer-specific driver gene alternations, making precision medicine possible. However, accurate genetic testing requires a sufficient amount of tumor cells in the specimen. The evaluation of tumor content ratio (TCR) from hematoxylin and eosin (H&E)-stained images has been found to vary between pathologists, making it an important challenge to obtain an accurate TCR. In this study, three pathologists exhaustively labeled all cells in 41 regions from 41 lung cancer cases as either tumor, non-tumor or indistinguishable, thus establishing a “gold standard” TCR. We then compared the accuracy of the TCR estimated by 13 pathologists based on visual assessment and the TCR calculated by an AI model that we have developed. It is a compact and fast model that follows a fully convolutional neural network architecture and produces cell detection maps which can be efficiently post-processed to obtain tumor and non-tumor cell counts from which TCR is calculated. Its raw cell detection accuracy is 92% while its classification accuracy is 84%. The results show that the error between the gold standard TCR and the AI calculation was significantly smaller than that between the gold standard TCR and the pathologist’s visual assessment (p<0.05). Additionally, the robustness of AI models across institutions is a key issue and we demonstrate that the variation in AI was smaller than that in the average of pathologists when evaluated by institution. These findings suggest that the accuracy of tumor cellularity assessments in clinical workflows is significantly improved by the introduction of robust AI models, leading to more efficient genetic testing and ultimately to better patient outcomes.

Experimental Study of Interference in Calcitonin Testing: A Case Report and Literature Review.

Calcitonin (CT) is a sensitive serum marker of medullary thyroid carcinoma usually detected via immunoassays; however, its levels are easily disturbed by several endogenous factors. The study aimed to discuss a case of suspected interference resulting in aberrant CT values and review previous reports of CT interference. A female patient visited our clinic with a physical ultrasound examination showing a slightly enlarged thyroid gland with small nodules. She had elevated CT levels, inconsistent with the clinical presentation and other findings. We evaluated the results by retesting using the same platform, platform validation, multiplex dilution, Polyethylene Glycol (PEG) precipitation, heterophilic blocking tubes, and RET gene analysis. Retesting CT using the same platform confirmed the high value obtained. However, serial dilution of the sample produced nonlinear results, suggesting some interference. While PEG precipitation did not significantly reduce the CT level, incubating the sample in HBTs normalized the CT value, indicating interference from heterophilic antibodies. Gene sequencing revealed no RET mutations. In cases where elevated CT levels are inconsistent with clinical presentations and other findings, the laboratory technicians should communicate with clinicians, analyze the reasons for the inconsistent results, and use different methods to verify the results. Accurate testing provides realistic and reliable data for doctors and patients and helps to avoid unnecessary procedures.

DROPLET MOTION ON HYDROPHOBIC AND HYDROPHILIC SURFACES USING LATTICE BOLTZMANN METHOD WITH BLOOD TEST MICROPAD APPLICATION

Laboratory sciences and disease diagnostics need accelerated medical tests, ease of testing for patients, and lower diagnostic costs. In this respect, one of the most innovative methods is paper-based microfluidic analysis tools, called micro-PAD. Since the chemical-based micro-PADs need sufficient blood to show an accurate blood test, the shape of the micro-PADs, the blood quantity, and the droplet flow to reach the end of the microchannels are important. Therefore, this research showed the liquid droplet motion on a heterogeneous (hydrophobic and hydrophilic) surface using the lattice Boltzmann method, and two star-shaped micro-PADs were modeled to investigate the effect of the droplet size and the droplet deflection from the center of the micro-PAD on liquid flow in microchannels. Based on the obtained results, fluid behavior was similar in two patterns (ten-pointed pattern and four-pointed pattern). This study compared the effects of parallel, convergent, and divergent channels in a four-pointed pattern on fluid velocity. The off-center error of the droplet was considered at the initial moment. The obtained results confirmed the direct relationship between the increased deviation of the center and the lack of uniformity in the fluid penetration. The present simulation results were validated against the laboratory data.

Sliding motility of Bacillus cereus mediates vancomycin pseudo-resistance during antimicrobial susceptibility testing.

The glycopeptide vancomycin is the antimicrobial agent-of-choice for the treatment of severe non-gastrointestinal infections with members of Bacillus cereus sensu lato (s.l.). Recently, sporadic detection of vancomycin-resistant phenotypes emerged, mostly for agar diffusion testing such as the disc diffusion method or gradient test (e.g. Etest®) method. In this work, we were able to disprove a preliminarily assumed high resistance to vancomycin in an isolate of B. cereus s.l. using broth microdilution and agar dilution. Microscopic imaging during vancomycin susceptibility testing showed spreading towards the inhibition zone, which strongly suggested sliding motility. Furthermore, transcriptomic analysis using RNA-Seq on the nanopore platform revealed several key genes of biofilm formation (e.g. calY, tasA, krsEABC) to be up-regulated in pseudo-resistant cells, substantiating that bacterial sliding is responsible for the observed mobility. Down-regulation of virulence (e.g. hblABCD, nheABC, plcR) and flagellar genes compared with swarming cells also confirmed the non-swarming phenotype of the pseudo-resistant isolate. The results highlight an insufficiency of agar diffusion testing for vancomycin susceptibility in the B. cereus group, and reference methods like broth microdilution are strongly recommended. As currently no guideline mentions interfering phenotypes in antimicrobial susceptibility testing of B. cereus s.l., this knowledge is essential to obtain reliable results on vancomycin susceptibility. In addition, this is the first report of sliding motility undermining accurate antimicrobial susceptibility testing in B. cereus s.l. and may serve as a basis for future studies on bacterial motility in susceptibility testing and its potential impact on treatment efficacy.

Summary of Research on Highway Bridge Vehicle Force Identification

Vehicle force identification is one of the core technical problems to be solved urgently in the management of transportation infrastructure, and it has also been a research hotspot in recent years. To promote the application of vehicle force identification technology on bridges and explore its development direction, the development status of indirect vehicle force identification methods based on bridge response is reviewed during this study. The basic theories of two major methods, including bridge weigh-in-motion (BWIM) and moving force identification (MFI), are described in detail in this study, and then, the key technical principles of bridge force identification are revealed. Secondly, the development status of BWIM in recent years is reviewed from three aspects, including test accuracy, applicability and test efficiency. Combined with a variety of theories, the current status of MFI is analyzed from the establishment of the solution to the equation. Finally, the development direction of an artificial neural network and machine vision technology are prospected in this study. The BP neural network has good self-learning ability and self-adaptive ability, but the algorithm needs to be improved. The identification method based on machine vision represents the current development direction in vehicle force identification, with great potential.

Predictive machine learning model for mechanical dilatation in transvenous lead extraction procedures

Abstract Background Transvenous lead extraction (TLE) remains a procedure that requires a high level of expertise, with a doubled risk of death and clinical failure when performed in low-volume centers compared to high-volume ones (1). Purpose The aim of this study was to create a machine learning (ML)-based risk stratification system for predicting the need for mechanical dilatation in patients undergoing TLE due to infection. Methods We designed a ML-based risk stratification system trained with data from our registry to predict the need for mechanical dilatation in patients undergoing TLE for infection. An extensive evaluation of 5 different ML models (k-nearest neighbors, support vector machine, decision tree, and decision tree ensembles, such as random forest and gradient boosting machine) was conducted to identify a classifier with the highest potential to correctly predict previously unseen patients. Results Data to train the model was extracted from our 25-year registry of patients undergoing TLE (June 1998 – March 2023), for a total of 491 patients (77.8% male; age 69.7 ± 12.8 years) and 938 leads (ICD 21.2%; pacing 78.8%; indwelling time 61 ± 60 months) removed with success in 100% of cases. Each patient was represented by a set of 21 attributes (14 clinical, 7 device-related). Manual traction (MT) was used in 27.5% of cases, and mechanical dilatation (MD) was employed in the remaining 72.5% of cases. 5-fold nested cross validation was used to estimate performances: in turn, 393 patients were used for training and model selection, and 98 patients were used for independent testing. According to the evaluation, Gradient Boosting Machine performed best, achieving test accuracy of 89% (+/- 2% std. dev.), test sensitivity of 95% (+/- 3% std. dev.), test specificity of 73% (+/- 8% std. dev.), test AUROC of 92% (+/- 1% std. dev.). A further interpretability analysis on the best performing decision tree was conducted, showing remarkable adherence between the internal decisions taken by the model to make predictions and the current clinical practice for TLE. Conclusion ML models have enhanced the prediction of MD requirements in TLE procedures. This data may support the referral of patients to specialized and high-volume centers, thereby improving preoperative risk assessment.Feature importanceROC Curves

Accelerated construction of stress relief music datasets using CNN and the Mel-scaled spectrogram.

Listening to music is a crucial tool for relieving stress and promoting relaxation. However, the limited options available for stress-relief music do not cater to individual preferences, compromising its effectiveness. Traditional methods of curating stress-relief music rely heavily on measuring biological responses, which is time-consuming, expensive, and requires specialized measurement devices. In this paper, a deep learning approach to solve this problem is introduced that explicitly uses convolutional neural networks and provides a more efficient and economical method for generating large datasets of stress-relief music. These datasets are composed of Mel-scaled spectrograms that include essential sound elements (such as frequency, amplitude, and waveform) that can be directly extracted from the music. The trained model demonstrated a test accuracy of 98.7%, and a clinical study indicated that the model-selected music was as effective as researcher-verified music in terms of stress-relieving capacity. This paper underlines the transformative potential of deep learning in addressing the challenge of limited music options for stress relief. More importantly, the proposed method has profound implications for music therapy because it enables a more personalized approach to stress-relief music selection, offering the potential for enhanced emotional well-being.