Training Time Research Articles

FedNIC: enhancing privacy-preserving federated learning via homomorphic encryption offload on SmartNIC

Federated learning (FL) has emerged as a promising paradigm for secure distributed machine learning model training across multiple clients or devices, enabling model training without having to share data across the clients. However, recent studies revealed that FL could be vulnerable to data leakage and reconstruction attacks even if the data itself are never shared with another client. Thus, to resolve such vulnerability and improve the privacy of all clients, a class of techniques, called privacy-preserving FL, incorporates encryption techniques, such as homomorphic encryption (HE), to encrypt and fully protect model information from being exposed to other parties. A downside to this approach is that encryption schemes like HE are very compute-intensive, often causing inefficient and excessive use of client CPU resources that can be used for other uses. To alleviate this issue, this study introduces a novel approach by leveraging smart network interface cards (SmartNICs) to offload compute-intensive HE operations of privacy-preserving FL. By employing SmartNICs as hardware accelerators, we enable efficient computation of HE while saving CPU cycles and other server resources for more critical tasks. In addition, by offloading encryption from the host to another device, the details of encryption remain secure even if the host is compromised, ultimately improving the security of the entire FL system. Given such benefits, this paper presents an FL system named FedNIC that implements the above approach, with an in-depth description of the architecture, implementation, and performance evaluations. Our experimental results demonstrate a more secure FL system with no loss in model accuracy and up to 25% in reduced host CPU cycle, but with a roughly 46% increase in total training time, showing the feasibility and tradeoffs of utilizing SmartNICs as an encryption offload device in federated learning scenarios. Finally, we illustrate promising future study and potential optimizations for a more secure and privacy-preserving federated learning system.

Wearable Activity Trackers and Physical Activity Levels Among Members of the Athens Medical Association in Greece.

Wearable Activity Trackers (WATs) offer real-time feedback on activity levels. We assessed the impact of WAT usage on physicians' exercise habits. Physicians from the Athens Medical Association, Greece (n = 742) responded to a self-administered questionnaire evaluating usage of WAT, demographic characteristics, specialty, and physical exercise habits. WHO guidelines recommend at least 150 min/week of moderate-intensity exercise in all healthy adults. Subjects were divided in Users of WATs (Group A), and Non-Users of WATs (Group B). This is an observational, cross-sectional study. There was no difference in baseline characteristics between the two groups (age, sex, body mass index). WATs were used by 38%. Between Group A and B, there was difference in mean exercise training time (302 ± 304 min vs. 210 ± 268 min, p < 0.001), higher percentage of WHO goal achievement (66.3% vs. 50.7%, p < 0.001), and greater awareness of WHO Guidelines (59.9% vs. 47.4%, p < 0.001). WATs were mostly used by four main specialties, with higher use from Cardiologists: Cardiology (47%), Endocrinology (44%), Surgery (35%) and Internal Medicine (25%), with a p = 0.045. Finally, users of WATs compared to non-users showed higher willingness to reduce body weight (58.5% vs. 48%, p = 0.01), apply dietary restrictions (36.5% vs. 29.6%, p = 0.05), and greater motivation for weekly physical exercise (74.1% vs. 32.4%, p < 0.001); Conclusion: Physicians using WATs demonstrate increased exercise training time, greater awareness of WHO guidelines and a higher propensity to implement dietary restrictions compared to non-users. Variations in WAT usage across medical specialties emphasize the need for targeted interventions to promote physical activity and enhance healthcare professionals' health.

Scene Classification of Remote Sensing Image Based on Multi-Path Reconfigurable Neural Network

Land image recognition and classification and land environment detection are important research fields in remote sensing applications. Because of the diversity and complexity of different tasks of land environment recognition and classification, it is difficult for researchers to use a single model to achieve the best performance in scene classification of multiple remote sensing land images. Therefore, to determine which model is the best for the current recognition classification tasks, it is often necessary to select and experiment with many different models. However, finding the optimal model is accompanied by an increase in trial-and-error costs and is a waste of researchers’ time, and it is often impossible to find the right model quickly. To address the issue of existing models being too large for easy selection, this paper proposes a multi-path reconfigurable network structure and takes the multi-path reconfigurable residual network (MR-ResNet) model as an example. The reconfigurable neural network model allows researchers to selectively choose the required modules and reassemble them to generate customized models by splitting the trained models and connecting them through modules with different properties. At the same time, by introducing the concept of a multi-path input network, the optimal path is selected by inputting different modules, which shortens the training time of the model and allows researchers to easily find the network model suitable for the current application scenario. A lot of training data, computational resources, and model parameter experience are saved. Three public datasets, NWPU-RESISC45, RSSCN7, and SIRI-WHU datasets, were used for the experiments. The experimental results demonstrate that the proposed model surpasses the classic residual network (ResNet) in terms of both parameters and performance.

Introducing edge intelligence to smart meters via federated split learning

The ubiquitous smart meters are expected to be a central feature of future smart grids because they enable the collection of massive amounts of fine-grained consumption data to support demand-side flexibility. However, current smart meters are not smart enough. They can only perform basic data collection and communication functions and cannot carry out on-device intelligent data analytics due to hardware constraints in terms of memory, computation, and communication capacity. Moreover, privacy concerns have hindered the utilization of data from distributed smart meters. Here, we present an end-edge-cloud federated split learning framework to enable collaborative model training on resource-constrained smart meters with the assistance of edge and cloud servers in a resource-efficient and privacy-enhancing manner. The proposed method is validated on a hardware platform to conduct building and household load forecasting on smart meters that only have 192 KB of static random-access memory (SRAM). We show that the proposed method can reduce the memory footprint by 95.5%, the training time by 94.8%, and the communication burden by 50% under the distributed learning framework and can achieve comparable or superior forecasting accuracy to that of conventional methods trained on high-capacity servers.

Grasp with push policy for multi-finger dexterity hand based on deep reinforcement learning

When interacting with the external environment, dexterous hands face various challenges, including interference in cluttered environments and difficulty in accurately locating target objects. We observe that the human hand is often at ease in the face of complex unstructured environments: pushing objects that can rearrange clutter, locating target objects, and creating space for fingers. In addition, the gripping action complements the push by achieving the precise movement of unrelated objects. Taking inspiration from this, to enable manipulators to perform accurate grasping tasks in complex environments, we propose a new data-driven grasping approach for robotic dexterous hands: the push-fit grasping approach. This method combines human grasping with model-free deep reinforcement learning to realize the robot's collaborative grasping ability. An end-to-end conditional converter grasping network is first trained, which takes the visual point cloud input to the action output. We use DQN algorithms for strategy training, further enhanced by fine-tuning methods for real-world learning. The experimental results show that the push strategy we learned significantly improves the grasping performance, and the success rate of cooperative grasping is increased by 8 %. It is worth noting that this fine-tuning method significantly reduces the actual training time, with the real world training taking up only one-fifth of the time in the simulated world. Even in demanding scenarios characterized by confusion and complexity, it facilitates rapid learning. This study provides new insights for effectively addressing complex challenges involving multi-fingered dexterous hands and human-computer interaction. Our code is available in the github repository.

A novel hybrid CNN-GRU model for website-based user interface code generator

As software engineering techniques evolve, conventional methods become less efficient due to prolonged development cycles and repetitive tasks. There is a growing need for automated systems that can incorporate User Interface (UI) designs into code. This study proposes a novel Hybrid CNN-GRU model for automatically generate website-based UI codes from UI design images. The proposed method consists of a visual model and a language-decoder part. The visual model employs pretrained CNN models, including EfficientNetV2B0, Xception, and EfficientNetV2M, to extract features from images, whereas a GRU serves as the language model and decoder. An empirical study was conducted to evaluate the effectiveness of various hybrid CNN-GRU approaches for generating UI codes. The EfficientNetV2B0 GRU model demonstrated the fastest training time, whereas the Xception GRU model was the most robust for various data complexities. The results of this study will advance the field of automated code generation, simplify the front-end development process, and shorten the time and effort necessitated for UI code implementations.

Assessment of validity, reliability, and feasibility of OMERACT ultrasound knee osteoarthritis scores in Egyptian patients with primary knee osteoarthritis.

Ultrasound (US) can evaluate all joint components affected by knee osteoarthritis (KOA); however, standardized scoring of US-detected pathology is needed to improve its diagnostic and monitoring capabilities. To examine the validity, reliability, and feasibility of the Outcome Measures in Rheumatology (OMERACT) ultrasound scoring for KOA, comparing with clinical and radiography measures, using predefined cutoff values. This cross-sectional study included 75 Egyptian patients with primary KOA. All patients had Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) score, bilateral knee radiography, and ultrasonography. Inter-observer reliability of ultrasound was evaluated in 30 knees by another newly trained operator. Most of the OMERACT-US KOA scores showed significant associations with WOMAC clinical scores, except for femoral cartilage damage and effusion. The synovitis score was significantly associated with WOMAC-pain score (p-value 0.046), while medial meniscus extrusion (MME) and medial osteophytes were significantly associated with WOMAC-stiffness score (p-value 0.009 and 0.023, respectively). MME and synovitis were significantly associated with WOMAC-physical score (p-value 0.035 and 0.020, respectively). The ultrasound scores also showed a strong correlation with radiographic scoring. Inter-observer reliability ranged from moderate to excellent agreement (k = 0.58 to k = 0.83); it was highest for lateral osteophytes (k = 0.83), good agreement for synovitis (k = 0.72), any osteophytes (k = 0.71), damage of femoral cartilage (k = 0.70), and moderate agreement for medial osteophytes (k = 0.58) and MME (k = 0.59). OMERACT-US scoring system for KOA demonstrated validity, reliability, and feasibility for evaluating both structural and inflammatory components. Using cutoff values improved the scoring reliability for osteophytes and MME. Key Points • OMERACT-US scores provide a valid assessment of inflammatory and structural components of knee osteoarthritis. • The following changes may improve the performance of the OMERACT-US scores. a. The binary score for effusion and synovial hypertrophy can be omitted, as they have no added value. b. A semi-quantitative grading for effusion may capture the impact of effusion on clinical outcomes. c. Added cutoff values to score medial meniscal extrusion, osteophytes, and pathological effusion improved the respective scores' reliability. d. Applying the updated OMERACT definition of synovitis. • OMERACT-US scores are reliable to be used with a newly trained operator, particularly when cutoff values are included, and proper training time is provided. • The OMERACT-US score is feasible to be used in clinical practice, as the time taken to perform was short, even for a newly trained operator.

Efficient pretraining model based on multi-scale local visual field feature reconstruction for PCB CT image element segmentation

Element segmentation is a key step in nondestructive testing of printed circuit boards (PCBs) based on computed tomography (CT) technology. In recent years, there has been rapid development of self-supervised pretraining technology that can obtain general image features without labeled samples and then uses a small amount of labeled samples to solve downstream tasks, which has good potential in PCB element segmentation. At present, a masked image modeling (MIM) pretraining model has been initially applied in PCB CT image element segmentation. However, due to the small and regular size of PCB elements such as vias, wires, and pads, the global visual field has redundancy for a single-element reconstruction, which may damage the performance of the model. Based on this issue, we propose an efficient pretraining model based on multi-scale local visual field feature reconstruction for PCB CT image element segmentation (EMLR-seg). In this model, the teacher-guided MIM pretraining model is introduced into PCB CT image element segmentation for the first time, to our knowledge, and a multi-scale local visual field extraction (MVE) module is proposed to reduce redundancy by focusing on local visual fields. At the same time, a simple four-Transformer-blocks decoder is used. Experiments show that EMLR-seg can achieve 88.6% mIoU on the PCB CT image dataset we proposed, which exceeds 1.2% by the baseline model, and the training time is reduced by 29.6 h, a reduction of 17.4% under the same experimental condition, which reflects the advantage of EMLR-seg in terms of performance and efficiency.

A Study of Data Processing Methods for Non-Contact Multispectral Method of Blood Oxygen Saturation.

Regular monitoring of blood oxygenation is important for disease prevention and treatment. Image photoplethysmography (IPPG) technology is a non-contact physiological parameter detection technology, which has been widely used in blood oxygenation detection. However, traditional imaging devices still have issues such as low detection accuracy, narrower receiving spectral range. In this paper, we proposed two improved detection methods based on the dual-wavelength measurement principle, that is, dual-band IPPG signal ratio method and dual-band IPPG signal AC/DC method. To verify the effectiveness of the two methods, we used different heartbeat period IPPG signals as sample data sets, and combined PLS and RF algorithms for model training, thus obtaining the best data processing method. The experimental results showed that the dual-band IPPG signal AC/DC method can effectively reduce the model training time. This method meets the strong demand for non-contact blood oxygen measurement and provides a new measurement idea.

Plastic Constitutive Training Method for Steel Based on a Recurrent Neural Network

The deep learning steel plastic constitutive model training method was studied based on the recurrent neural network (RNN) model to improve the allocative efficiency of the deep learning steel plastic constitutive model and promote its application in practical engineering. Two linear hardening constitutive datasets of steel were constructed using the Gaussian stochastic process. The RNN, long short-term memory (LSTM), and gated recurrent unit (GRU) were used as models for training. The effects of the data pre-processing method, neural network structure, and training method on the model training were analyzed. The prediction ability of the model for different scale series and the corresponding data demand were evaluated. The results show that LSTM and the GRU are more suitable for stress–strain prediction. The marginal effect of the stacked neural network depth and number gradually decreases, and the hysteresis curve can be accurately predicted by a two-layer RNN. The optimal structure of the two models is A50-100 and B150-150. The prediction accuracy of the models increased with the decrease in batch size and the increase in training batch, and the training time also increased significantly. The decay learning rate method could balance the prediction accuracy and training time, and the optimal initial learning rate, batch size, and training batch were 0.001, 60, and 100, respectively. The deep learning plastic constitutive model based on the optimal parameters can accurately predict the hysteresis curve of steel, and the prediction abilities of the GRU are 6.13, 6.7, and 3.3 times those of LSTM in short, medium, and long sequences, respectively.

Concrete crack classification based on fourier image enhancement and convolutional neural network

This paper investigates the application of Fourier image enhancement combined with Convolutional Neural Networks (CNNs) for detecting cracks in concrete structures. Fourier enhancement is used to preprocess crack images, improving their clarity and reducing noise, which in turn enhances the performance of the CNN in accurately classifying cracks. The results demonstrate that this combination improves the classification accuracy, with the enhanced images achieving a higher accuracy compared to non-enhanced images. Additionally, the study examines the effects of this preprocessing on CNN training time. However, accuracy varies depending on the dataset used, with one dataset reaching a maximum accuracy of 95% after enhancement. These findings highlight the potential of using frequency-domain image enhancement techniques in conjunction with deep learning models for structural health monitoring.

Research on faults diagnosis of dynamic balance in scroll compressor based on CWT-MViTV2

Abstract The dynamic balancing of rotor systems is a critical factor affecting the vibration of scroll compressors. Traditionally, the detection and correction of rotor dynamic balancing systems have relied on time-domain and frequency-domain analysis of vibration signals, as well as validation using rotor dynamic balancing test platforms. This paper proposes a novel fault diagnosis method for the dynamic balancing of scroll compressor rotor systems. This method acquires the time-domain vibration signals of the scroll compressor through online detection experiments. These signals are then transformed into time-frequency maps using continuous wavelet transform(CWT). By integrating the MViTV2 neural network, this approach enables effective identification of dynamic balancing fault types without requiring disassembly or shutdown of the machine. The fault types, corresponding to the offset of the balancing weight’s centroid, will provide direct data support for subsequent validation work. This study compares various neural network models combined with time-frequency maps and demonstrates that the proposed model achieves the highest accuracy of 99.749% compared to other models, and the method's generalizability is validated in the public dataset. Furthermore, the proposed model maintains a high accuracy of 94.872% in high-noise environments. After improvement, the accuracy of the model has been increased to 95.641%, the training time and the diagnosis time has been reduced to 0.240 iter/s and 0.0277s.

Illustration of the Financial Impact of Usage of Initial Specimen Diversion Devices (ISDD) or Continued Educational Initiatives on the Costs from Blood Culture Contamination

Abstract Introduction/Objective Blood culture contamination (BCC) has the potential to lead to additional healthcare system costs and harm to patient. Two ways to reduce BCC have been explored in the literature and are economically modeled including initial specimen diversion devices (ISDD) and continued education on phlebotomy. Methods/Case Report A literature review on the economic costs of BCC, ISDD efficacy, and phlebotomy education to reduce BCC rates was undertaken. The reported average BCC at a veteran affairs medical center (VAMC), its reported rate during a month of phlebotomy training, the cost of an ISDD device at VAMC, and the average wage for 4 hours of phlebotomy manager time for training was noted. An illustrative economic model involving 100 patients was used at BCC rates of 1%, 2%, and 3% along with what had been observed at the VAMC was used for economic modeling. Results (if a Case Study enter NA) It was previously reported by Skoglund et al that in laboratories with rapid diagnostic testing modalities (such as MALDI-TOF or PCR testing), the cost per BCC event is $13,026 versus $8,287 in costs for a negative blood culture. ISDD efficacy in the literature reviewed (74.1% to 87.5%), continuous education efficacy (16.2% to 57.8%), ISDD device cost ($17.70 per device), and the hourly wage of a phlebotomy supervisor ($29.34) were noted. The reported average BCC rate from a regional VAMC was 1.33%, and during the month with education, the rate was 0.5%. From a model of 100 patients, ISDD device savings ranged from $1,741.60-$2,376.63 (1% BCR), $5,253.20-$6,523.25 (2% BCR), and $8,764.79-$10,669.88 (3% BCR). Education savings ranged from $755.98-$2,727.41 (1% BCR), $1,523.70-5,466.55 (2% BCR), and $2,291.42-$8,205.69 (3% BCR). The observed impact of education at the regional VAMC would have resulted in savings of $650.83 per 100 patients. Conclusion Both education and ISDD devices can result in significant savings from the costs of BCC.

Transfer learning-assisted modulation classification and OSNR estimation using constellation diagrams

Abstract A cost-effective deep transfer learning (TL) learned with characteristics derived from the constellation diagram is proposed for executing modulation classification and optical signal-to-noise ratio (OSNR) estimation for the future generation of elastic optical networks. The information acquired with the ImageNet dataset is transferred using pre-trained TL versions including VGG19, ResNet152V2, InceptionV3, and ResNet50 to identify various modulation formats and their corresponding OSNR. Adam optimizer is employed to ascertain the appropriate settings for the hyperparameters. Numerical results demonstrate that the proposed VGG19 method achieves the highest level of accuracy (100%) in the recognition of various modulation forms. To fulfil the requirements of actual use, OSNR monitoring is also explored, with a total precision equal to 95.2%. Furthermore, a thorough investigation of the impact of training data dimensions, on TL performance is conducted. The outcomes of the proposed method demonstrate that the suggested TL-based algorithms are more accurate and need much less training and testing time than non-TL approaches.

Suicide prevention safety planning in the US Department of defense: Qualitative assessment of training.

Suicide rates in the military are a significant public health concern. The suicide prevention safety planning intervention is a brief and effective intervention in which a provider and patient work collaboratively to recognize warning signs of a suicidal crisis and create a defined list of coping strategies and supports for use during future crises (Stanley & Brown, 2012). Implementation of safety planning has been supported by the Department of Defense (DoD), yet readily available training and continuing education for healthcare providers in this intervention is limited and passive in nature. Existing safety planning training experiences and needs of DoD behavioral health providers are unknown. The present study is an exploratory qualitative assessment of current safety planning in DoD to inform the design and development of an interactive virtual standardized training patient. Ten military behavioral health providers completed semi-structured interviews. Thematic content analysis was conducted, and three themes are described in this paper: variety of training received, barriers to receiving suicide safety planning training, and desires for future training. Lack of protected time for training and a lack of access to training resources were identified as key barriers. Rich data obtained can help inform the key design features and relevance of new safety planning intervention training approaches.

Heartbeat classification using various machine learning models: A comparative study

Cardiac arrhythmias, known as irregular heartbeats, pose a notable health threat that necessitates prompt diagnosis, as untreated arrhythmias can lead to severe heart complications. Among the various methods for arrhythmia detection, electrocardiography is the most prevalent due to its non-invasive monitoring of heart activity. However, manual electrocardiogram (ECG) analysis is inefficient and prone to errors, prompting the exploration of machine learning (ML) models for ECG feature recognition. Integrating ML models with ECG analysis can revolutionize cardiac diagnostics by improving healthcare efficiency and outcomes by enhancing the accuracy and consistency of existing approaches as well as their processing speed for large datasets. Unfortunately, current ML methods encounter two key limitations: prolonged training times and the need for manual feature selection. To address these issues, we propose using ML models enhanced with innovative techniques such as the Fourier transform (FT) and Gaussian noise injection for improved cardiac health assessment. To validate this approach, we utilized statistical tools, including Pearson correlation and p-values, to uncover relationships within the data. In addition, we employed the FT technique to extract and analyze frequency-domain features. Our comparative study of different ML models relied on metrics such as accuracy, precision, recall, F1 score, and receiver operating characteristic area under the receiver operating characteristic curve, demonstrating XGBoost&amp;rsquo;s impressive average recall of 0.956 with 99.96% overall accuracy. An average precision of 0.956 further underscored the accuracy of XGBoost&amp;rsquo;s predictions, indicating its high level of reliability in distinguishing various cardiac conditions. These results highlight the considerable potential of ML techniques for precise ECG-based clinical diagnoses, helping healthcare professionals make more accurate and timely decisions in patient care.

Artificial neural network based decentralized current-sharing control for parallel connected DC-DC converters in DC microgrid application

To create a non-interrupted, economical and reliable direct current (DC) microgrid, multiple DC-DC converters can be connected in parallel. Current sharing among these multiple converters becomes essential for proper operation of the system. This study proposes an artificial neural network (ANN) based control technique for parallel connected DC-DC boost converters which ensures accurate current sharing according to the specified maximum limits. Levenberg-Marquardt algorithm-based ANN network is used to reduce the training time and, also to achieve near ≈ 100 % accuracy of the training data. ANN control provides better voltage regulation, accurate current sharing unlike the conventional proportional-integral (PI) based control which faces issues such as inaccurate current sharing, high transient, peak overshoots and steady state error during sudden change in the system. The efficient functioning of the proposed control method is verified by simulating two parallel connected DC-DC converters using MATLAB/Simulink. A hardware prototype of converters rating ≈ 250 W using TMS320F28379D Digital signal processor controller is also developed to verify the performance and effectiveness of the ANN based control in comparison to PI based technique. The ANN based technique is faster to achieve the reference voltage, and the peak overshoot is approximately 75 % lesser than the PI based control.

MobileNet-CAL: Tomato pest and disease classification method based on transfer learning and attention mechanism

Based on the MobileNetV3 model, this paper introduces the Coordinate attention (CA) attention mechanism and transfer learning to propose the MobileNet-CAL model to solve the problems of insufficient attention to model channel information, long classification training time, high training cost, and weak generalization ability. The model combines data enhancement, attention mechanism and transfer learning technology to improve the classification effect of the model, and updates the fully connected layer at the same time. Finally, the model is optimized by combining the Auxiliary loss function and the ASGD optimizer to improve the classification accuracy of the model, speed up the collection speed, and enhance the robustness of the model. Finally, the model is evaluated from multiple angles such as Accuracy, F1-score, Precision, Recall, and Specificity in the tomato pest and disease leaf dataset. Its Accuracy reaches 98.59% an accuracy of, which is improved by, and by, 4.47%respectively, compared with the original model. Experiments show that MobileNet-CAL can improve the classification accuracy to a certain extent, verifying that the model proposed 4.05%、3.43%、16.57% in 4.13%this paper is effective in the classification task of tomato pest and disease leaves.

Malware Detection Based on API Call Sequence Analysis: A Gated Recurrent Unit–Generative Adversarial Network Model Approach

Malware remains a major threat to computer systems, with a vast number of new samples being identified and documented regularly. Windows systems are particularly vulnerable to malicious programs like viruses, worms, and trojans. Dynamic analysis, which involves observing malware behavior during execution in a controlled environment, has emerged as a powerful technique for detection. This approach often focuses on analyzing Application Programming Interface (API) calls, which represent the interactions between the malware and the operating system. Recent advances in deep learning have shown promise in improving malware detection accuracy using API call sequence data. However, the potential of Generative Adversarial Networks (GANs) for this purpose remains largely unexplored. This paper proposes a novel hybrid deep learning model combining Gated Recurrent Units (GRUs) and GANs to enhance malware detection based on API call sequences from Windows portable executable files. We evaluate our GRU–GAN model against other approaches like Bidirectional Long Short-Term Memory (BiLSTM) and Bidirectional Gated Recurrent Unit (BiGRU) on multiple datasets. Results demonstrated the superior performance of our hybrid model, achieving 98.9% accuracy on the most challenging dataset. It outperformed existing models in resource utilization, with faster training and testing times and low memory usage.

Classification of Infant Crying Sounds Using SE-ResNet-Transformer

Recently, emotion analysis has played an important role in the field of artificial intelligence, particularly in the study of speech emotion analysis, which can help understand one of the most direct ways of human emotional communication—speech. This study focuses on the emotion analysis of infant crying. Within cries lies a variety of information, including hunger, pain, and discomfort. This paper proposes an improved classification model using ResNet and transformer. It utilizes modified Mel-frequency cepstral coefficient Mel-frequency cepstral coefficient (MFCC) features obtained through feature engineering from infant cries and integrates SE attention mechanism modules into residual blocks to enhance the model’s ability to adjust channel weights. The proposed method achieved 93% accuracy rate in experiments, offering advantages of shorter training time and higher accuracy compared to other traditional models. It provides an efficient and stable solution for infant cry classification.