Training Approach Research Articles

The impact of multicentric datasets for the automated tumor delineation in primary prostate cancer using convolutional neural networks on 18F-PSMA-1007 PET

PurposeConvolutional Neural Networks (CNNs) have emerged as transformative tools in the field of radiation oncology, significantly advancing the precision of contouring practices. However, the adaptability of these algorithms across diverse scanners, institutions, and imaging protocols remains a considerable obstacle. This study aims to investigate the effects of incorporating institution-specific datasets into the training regimen of CNNs to assess their generalization ability in real-world clinical environments. Focusing on a data-centric analysis, the influence of varying multi- and single center training approaches on algorithm performance is conducted.MethodsnnU-Net is trained using a dataset comprising 161 18F-PSMA-1007 PET images collected from four distinct institutions (Freiburg: n = 96, Munich: n = 19, Cyprus: n = 32, Dresden: n = 14). The dataset is partitioned such that data from each center are systematically excluded from training and used solely for testing to assess the model's generalizability and adaptability to data from unfamiliar sources. Performance is compared through a 5-Fold Cross-Validation, providing a detailed comparison between models trained on datasets from single centers to those trained on aggregated multi-center datasets. Dice Similarity Score, Hausdorff distance and volumetric analysis are used as primary evaluation metrics.ResultsThe mixed training approach yielded a median DSC of 0.76 (IQR: 0.64–0.84) in a five-fold cross-validation, showing no significant differences (p = 0.18) compared to models trained with data exclusion from each center, which performed with a median DSC of 0.74 (IQR: 0.56–0.86). Significant performance improvements regarding multi-center training were observed for the Dresden cohort (multi-center median DSC 0.71, IQR: 0.58–0.80 vs. single-center 0.68, IQR: 0.50–0.80, p < 0.001) and Cyprus cohort (multi-center 0.74, IQR: 0.62–0.83 vs. single-center 0.72, IQR: 0.54–0.82, p < 0.01). While Munich and Freiburg also showed performance improvements with multi-center training, results showed no statistical significance (Munich: multi-center DSC 0.74, IQR: 0.60–0.80 vs. single-center 0.72, IQR: 0.59–0.82, p > 0.05; Freiburg: multi-center 0.78, IQR: 0.53–0.87 vs. single-center 0.71, IQR: 0.53–0.83, p = 0.23).ConclusionCNNs trained for auto contouring intraprostatic GTV in 18F-PSMA-1007 PET on a diverse dataset from multiple centers mostly generalize well to unseen data from other centers. Training on a multicentric dataset can improve performance compared to training exclusively with a single-center dataset regarding intraprostatic 18F-PSMA-1007 PET GTV segmentation. The segmentation performance of the same CNN can vary depending on the dataset employed for training and testing.

Executive functions and divergent thinking in young adults: Evidence from individual differences and cognitive training studies

The interplay between executive functions and divergent thinking remains a debated question in creativity research. To address this issue, two studies were conducted from different perspectives. The first study explored which of the executive functioning components—namely working memory, inhibition, and shifting—predicted divergent thinking. A sample of 203 college students performed a battery of tasks assessing divergent thinking, executive functions, and related cognitive variables (fluid intelligence, verbal fluency, and processing speed). After statistically controlling for these variables, only working memory emerged as a significant predictor of creativity scores in the Alternative Uses Task, while inhibition and shifting did not. In the second study, a cognitive training approach was employed to further elucidate the potential efficacy of working memory training in fostering divergent thinking. The results found that the experimental group, which received a 15-day dual n-back task training, demonstrated greater improvements in the creativity score of the Alternative Uses Task compared to the control group, suggesting a transfer effect from working memory training to divergent thinking. This research underscores the plasticity of divergent thinking and highlights a plausible avenue for enhancing creativity through targeted cognitive training.

Maintenance cost optimisation in critical single-component systems: A technician’s training approach based on a joint learning–forgetting and fuzzy maintenance quality model

Effective maintenance in critical systems is crucial to ensure their operability and safety. Maintenance technicians play an important role in achieving the safety and reliability objectives of critical system components. However, the role of the technical team in maintenance quality is often overlooked. Our study addresses this knowledge gap by developing a methodology to optimise the training scheduling of maintenance technicians in single-component systems, focusing on a learn-and-forget approach. Our main objective is to improve maintenance efficiency by mitigating the negative impact of technician expertise level. To this end, we implemented a simulation-based optimisation, modelling the evolution of technician experience and its influence on maintenance costs. As a case study, we used an urban aerial passenger ropeway system, demonstrating that proper training of technical staff can significantly reduce both corrective and preventive maintenance costs. Our methodology identified three strategic areas for training technical staff throughout the system lifecycle. The results highlight the importance of investing in the training scheduling of technical equipment in critical systems and suggest that an appropriate training strategy can improve operational efficiency and reduce maintenance costs. In conclusion, our research underlines the need to actively consider the human factor in the maintenance of critical systems to ensure their optimal performance and long-term safety.

Enhancing the competencies of obstetrical nurses and midwives in high-risk pregnancy management through simulation-based training in Lao people's democratic republic: A pilot study

BackgroundSimulation-based training has been widely used as a valuable strategy for learning and evaluating clinical skills at different levels of nursing and midwifery education. The impact of simulation training on intensive management for high-risk pregnancy in a low-resource setting has been limited. AimTo examine the effect of simulation-based training with low-fidelity mannequins on obstetrical nurses and midwives’ knowledge, attitude, and skills for high-risk pregnancy management in a low-resource setting. MethodDuring September 2023, twenty-five obstetrical nurses or midwives who worked in five tertiary public hospitals in Vientiane Prefecture participated in the three-day training workshops for intensive management in high-risk pregnant women and newborns that used a simulation-based training approach integrating problem-based learning. The evaluated criteria of knowledge, attitudes, and skills pre- and post-test scores were statistically compared. FindingsWorkshop trainees demonstrated an increase significantly in knowledge for high-risk pregnancy management (p = 0.012), attitude toward high-risk pregnancy management (p = 0.000), and attitude toward simulation-based training design (p = 0.002). The clinical skills were used on the simulation performance checklist, and the pre-posttest gain in overall performance scores had a statistically significant difference (p = 0.000). The mean score of postpartum hemorrhage management skills was 11.48±2.23, which increased the highest score among all skills. ConclusionsThe simulation-based training in high-risk pregnancy management improves the knowledge, attitude, and skills of nurses and midwives in low-resource settings. Next steps include direct observation of trainees in the clinical setting to assess their competence in ensuring patient safety, achieving positive pregnancy outcomes, and enhancing satisfaction.

Optimising orthodox style learning in physical education through a variational training model

Background: The research problem is to improve our understanding of how different training approaches can improve bullet-shot results. Research Objectives: This study aims to investigate the impact of different training models on the learning outcomes of orthodox-style shot put. Methods: The research method employed was an experiment with a one-group pre-test and post-test design to enhance student learning outcomes using a varied training model. The subjects consisted of 32 students from a high school in Palembang. The study was conducted over 30 days, with four weekly meetings. During each meeting, 2-3 training models were presented, resulting in a total of 10 variation training models, namely Passing Techniques in Football Baseball, Techniques that Hold the Repulsion, Hanging the Ball, Resisting Pull, Clap Hand, Reject Target, Throwing Darts, Shoot in Target, and Load Rejection One Shoot. Data analysis used SPSS version 23 to calculate the significant value of the pretest and posttest results. Finding/Results: The study found that using varied training models resulted in significantly better learning outcomes in orthodox-style bullet throwing. Conclusion: The study concluded that varied training models have a considerable positive influence on learning outcomes in the orthodox style of instruction. This suggests that the findings contribute to a deeper understanding of effective training methods in sports education and have practical implications for athletic training programs. We could conduct longitudinal studies to investigate the long-term impact of diverse training on skill retention and entry into competitive athletics.

Parotid Gland Segmentation Using Purely Transformer-Based U-Shaped Network and Multimodal MRI.

Parotid gland tumors account for approximately 2% to 10% of head and neck tumors. Segmentation of parotid glands and tumors on magnetic resonance images is essential in accurately diagnosing and selecting appropriate surgical plans. However, segmentation of parotid glands is particularly challenging due to their variable shape and low contrast with surrounding structures. Recently, deep learning has developed rapidly, and Transformer-based networks have performed well on many computer vision tasks. However, Transformer-based networks have yet to be well used in parotid gland segmentation tasks. We collected a multi-center multimodal parotid gland MRI dataset and implemented parotid gland segmentation using a purely Transformer-based U-shaped segmentation network. We used both absolute and relative positional encoding to improve parotid gland segmentation and achieved multimodal information fusion without increasing the network computation. In addition, our novel training approach reduces the clinician's labeling workload by nearly half. Our method achieved good segmentation of both parotid glands and tumors. On the test set, our model achieved a Dice-Similarity Coefficient of 86.99%, Pixel Accuracy of 99.19%, Mean Intersection over Union of 81.79%, and Hausdorff Distance of 3.87. The purely Transformer-based U-shaped segmentation network we used outperforms other convolutional neural networks. In addition, our method can effectively fuse the information from multi-center multimodal MRI dataset, thus improving the parotid gland segmentation.

Extensive Multilabel Classification of Brain MRI Scans for Infarcts Using the Swin UNETR Architecture in Deep Learning Applications.

To distinguish infarct location and type with the utmost precision using the advantages of the Swin UNEt TRansformers (Swin UNETR) architecture. The research employed a two-phase training approach. In the first phase, the Swin UNETR model was trained using the Ischemic Stroke Lesion Segmentation Challenge (ISLES) 2022 dataset, which included cases of acute and subacute infarcts. The second phase involved training with data from 309 patients. The 110 categories result from classifying infarcts based on 22 specific brain regions. Each region is divided into right and left sides, and each side includes four types of infarcts (acute, acute lacunar, subacute, subacute lacunar). The unique architecture of Swin UNETR, integrating elements of both the transformer and u-net designs with a hierarchical transformer computed with shifted windows, played a crucial role in the study. During Swin UNETR training with the ISLES 2022 dataset, batch loss decreased to 0.8885±0.1897, with training and validation dice scores reaching 0.4224±0.0710 and 0.4827±0.0607, respectively. The optimal model weight had a validation dice score of 0.5747. In the patient data model, batch loss decreased to 0.0565±0.0427, with final training and validation accuracies of 0.9842±0.0005 and 0.9837±0.0010. The results of this study surpass the accuracy of similar studies, but they involve the issue of overfitting, highlighting the need for future efforts to improve generalizability. Such detailed classifications could significantly aid physicians in diagnosing infarcts in clinical settings.

SEMG-driven Hand Dynamics Estimation with Incremental Online Learning on a Parallel Ultra-Low-Power Microcontroller.

Surface electromyography (sEMG) is a State-of-the-Art (SoA) sensing modality for non-invasive human-machine interfaces for consumer, industrial, and rehabilitation use cases. The main limitation of the current sEMG-driven control policies is the sEMG's inherent variability, especially cross-session due to sensor repositioning; this limits the generalization of the Machine/Deep Learning (ML/DL) in charge of the signal-to-command mapping. The other hot front on the ML/DL side of sEMG-driven control is the shift from the classification of fixed hand positions to the regression of hand kinematics and dynamics, promising a more versatile and fluid control. We present an incremental online-training strategy for sEMG-based estimation of simultaneous multi-finger forces, using a small Temporal Convolutional Network suitable for embedded learning-on-device. We validate our method on the HYSER dataset, cross-day. Our incremental online training reaches a cross-day Mean Absolute Error (MAE) of (9.58 ± 3.89)% of the Maximum Voluntary Contraction on HYSER's RANDOM dataset of improvised, non-predefined force sequences, which is the most challenging and closest to real scenarios. This MAE is on par with an accuracy-oriented, non-embeddable offline training exploiting more epochs. Further, we demonstrate that our online training approach can be deployed on the GAP9 ultra-low power microcontroller, obtaining a latency of 1.49 ms and an energy draw of just 40.4 uJ per forward-backward-update step. These results show that our solution fits the requirements for accurate and real-time incremental training-on-device.

Driving competence in older age: psychological assessment and interventions

The article introduces the issue of psychological assessment and training for driving safety, driving competence, and the mobility of older drivers. The psychological methods for assessing driving competence and the training approaches for maintaining mobility in old age are outlined and illustrated with examples. Preserving driving ability in old age remains atask for which development needs and research gaps must be specified. Although periodic mandatory examinations based on age do not yield any relevant improvement in enhancing traffic safety, they do not cause harm and are widely accepted. It is desirable to move away from atheory focusing on deficits of older age and instead develop an assessment of the comprehensive functions, abilities, and resources of elderly drivers to help preserve the mobility and participation of that specific demographic in traffic.

Refining CycleGAN with attention mechanisms and age-Aware training for realistic Deepfakes

In the evolving landscape of deep learning technologies, the emergence of Deepfakes and synthetic media is becoming increasingly prominent within digital media production. This research addresses the limitations inherent in existing face image generation algorithms based on Generative Adversarial Networks (GAN), particularly the challenges of domain irrelevancy and inadequate facial detail representation. The study introduces an enhanced face image generation algorithm, aiming to refine the CycleGAN framework. The enhancement involves a two-fold strategy: firstly, the generator's architecture is refined through the integration of an attention mechanism and adaptive residual blocks, enabling the extraction of more nuanced facial features. Secondly, the discriminator's accuracy in distinguishing real from synthetic images is improved by incorporating a relative loss concept into the loss function. Additionally, this study presents a novel model training approach that incorporates age constraints, thereby mitigating the effects of age variations on the synthesized images. The effectiveness of the proposed algorithm is empirically validated through comparative analysis with existing methodologies, utilizing the CelebA dataset. The results demonstrate that the proposed algorithm significantly enhances the realism of generated face images, outperforming current methods in terms of Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM), while also achieving notable improvements in subjective visual quality. The implementation of this advanced method is anticipated to substantially elevate the efficiency and quality of digital media production, contributing positively to the broader field of digital media creation.

Db2une: Tuning Under Pressure via Deep Learning

Modern database systems including IBM Db2 have numerous parameters, "knobs," that require precise configuration to achieve optimal workload performance. Even for experts, manually "tuning" these knobs is a challenging process. We present Db2une, an automatic query-aware tuning system that leverages deep learning to maximize performance while minimizing resource usage. Via a specialized transformer-based query-embedding pipeline we name QBERT, Db2une generates context-aware representations of query workloads to feed as input to a stability-oriented, on-policy deep reinforcement learning model. In Db2une, we introduce a multi-phased, database meta-data driven training approach---which incorporates cost estimates, interpolation of these costs, and database statistics---to efficiently discover optimal tuning configurations without the need to execute queries. Thus, our model can scale to very large workloads, for which executing queries would be prohibitively expensive. Through experimental evaluation, we demonstrate Db2une's efficiency and effectiveness over a variety of workloads. We compare it against the state-of-the-art query-aware tuning systems and show that the system provides recommendations that surpass those of IBM experts.

Analysis of the relationship between attitudes, training, and perceptions of food handlers on supervision and the implementation of personal hygiene among food handlers

Background: Poor personal hygiene can affect the cleanliness of food provided to patients. A preliminary study conducted at XX Hospital Nutrition Installation found that five food handlers failed to use complete personal protective equipment (PPE), and two food handlers engaged in conversation while preparing food. This research aims to identify factors influencing the implementation of personal hygiene among food handlers at XX Hospital Nutrition Installation. Method: This quantitative study employed a cross-sectional design from November 2019 to March 2020. The sample included 43 food handlers selected through total sampling. Data analysis utilized SPSS and Chi-square tests. Results: The study indicated that 48.8% of food handlers exhibited inadequate personal hygiene, 41.9% had limited knowledge, 53.5% demonstrated negative attitudes, 60.5% did not receive training, and 51.2% perceived inadequate supervision from their superiors. Significant correlations were observed between knowledge (p=0.004), attitude (p=0.000), training (p=0.000), and perception of supervision (p=0.004) with the implementation of personal hygiene. Conclusion: Attitude, training, and perception of supervision significantly influence the adherence to personal hygiene among food handlers at XX Hospital 's Nutrition Installation. It is recommended that the Nutrition Installation enhance supervision of food handlers' personal hygiene and conduct certified training programs on food hygiene and sanitation. Novelty/Originality: This study provides new insights into the relationship between knowledge, attitudes, training, and perceptions of supervision with implementing personal hygiene in the context of hospital food handlers. The findings suggest that although multiple factors influence individual hygiene, a better training and supervision-based approach can significantly improve compliance with hygiene practices, which is an essential contribution to managing food quality in the hospital environment.

Employee risk recognition and reporting of malicious elicitations: longitudinal improvement with new skills-based training.

Numerous security domains would benefit from improved employee risk recognition and reporting through effective security training. This study assesses the effectiveness of a new skills-based training approach to improve risk recognition and reporting of malicious elicitations. Malicious elicitations are techniques that strategically use conversation (i.e., online, in writing, in person, or over the phone) with the sole purpose of collecting sensitive, non-publicly available information about business operations, people, or technological assets without raising suspicion. To an untrained observer, a skilled elicitor can make conversations seem analogous to many professional networking situations such as those experienced over email and at conferences. A 12-month longitudinal experimental study was conducted to test training effectiveness on employees of a large corporation that focuses on serving national security needs and the public interest. Half of participants were randomly assigned to receive traditional awareness-based training (i.e., reviewing informational slides) while the other half of participants received a new skills-based training that allowed them-over the course of five weeks-to iteratively practice skills learned in the training and receive feedback on their performance in their day-to-day work environment. Following training for both experimental groups, malicious elicitations and benign professional networking test messages were sent (via email & text message) to unaware employee participants for 12 months. Findings revealed that skills-based training improved reporting of malicious elicitations and lasted for up to 12 months compared to traditional awareness-based training.

A unified benchmark for deep reinforcement learning-based energy management: Novel training ideas with the unweighted reward

Deep reinforcement learning stands as a powerful force in the realm of intelligent control for hybrid power systems, yet some imperfections persist in the positive progression of learning-based strategies, necessitating the proposal of essential solutions to address these flaws. Firstly, a public and reliable benchmark model for hybrid powertrains and the optimization results of energy management strategies are essential. Hence, two Python-based standard deep reinforcement learning agents and four Simulink-based hybrid powertrains are employed, forming a co-simulation training approach as the reliable solution. Secondly, a detailed analysis from the perspectives of range, magnitude, and importance reveals that the optimization terms in traditional reward functions can mislead the agent during the training process and require cumbersome weight tuning. Accordingly, this paper proposes a novel training idea that combines the rule-based engine start-stop with an unweighted reward tailored for optimizing engine efficiency and facilitating training progress. Finally, a hardware-in-the-loop test is performed, treating the P2 hybrid electric vehicle as the target. The results show that two deep reinforcement learning-based energy management strategies achieved fuel economies of 6.537 L/100 km and 6.330 L/100 km, respectively, and more efficient and reasonable control sequences ensure the working state of the engine as well as the state of charge of batteries.

Enhancing eye-hand coordination in older adults: A nature-inspired drone training approach

Enhancing eye-hand coordination in older adults: A nature-inspired drone training approach

Real Classroom vs. Virtual Classroom: Volunteer Subtitle Translators’ Website as an Example of a Social Constructivist Approach to Translator Education

This descriptive study explores an interrelation between online volunteer subtitle organization and a social constructivist approach in translation training from a pedagogical point of view. The study sets out to test the hypothesis that technology-enabled online volunteer translators’ communities stand as an example of a social constructivist approach to translator training in a virtual context. Through an analysis of the forum section of the website as well as the workflow, similar characteristics of both the online platform and a Social Constructivist Approach in translator training by Kiraly are extracted. The study thus aims to identify if volunteer translator organizations can be an example of a Social Constructivist Approach to translator training in subtitle translation. The case study focuses on an online volunteer translators’ website named www.turkcealtyazi.org. The empirical data collected is based on the analysis of selected samples from the forum section of the website. The data derived is then compared with the model of the social constructivist approach in translator education to discover the similarities in terms of the relations between members, the workflow, the distribution of tasks, etc. The findings map how the online volunteer organization is similar to a social constructivist model applied in translator training in a classroom setting.

Professional development approaches and business educators’ job performance in tertiary institutions in Imo State

The study examined professional development approaches and business educators’ job performance in tertiary institutions in Imo state. It specifically focused on the relationship between in-service and on-the-job training approaches of professional development and the teaching and research performance of business educators in tertiary institutions. A co-relational research design was adopted in the study. The population of the study comprised 52 business educators in the two tertiary institutions currently running business education programme in Imo state. Data was collected from the respondents with the means of a structured questionnaire designed by the researchers. Data collected for the study was analyzed with mean (X) and Pearson Product Moment Correlation Coefficient (r). Analysis of the data collected revealed that both in-service and on-the-job training have a very strong positive relationship with business educators’ job performance in tertiary institutions. Consequently, it was concluded that in-service and on-the-job training approaches of professional development play important roles in the performance of business educators in tertiary institutions. Based on the findings of the study, it was recommended that: government should come up with a compulsory policy that will coerce management of institutions to prioritize regular professional development of their employees; and Management of institutions should combine both in-service and on-the-job training approaches in the professional development of their employees.

DEVELOPMENT OF AN ELECTRONIC EXAMINATION PLATFORM USING FACE RECOGNITION METHODS

Online systems face a major challenge in efficiently monitoring participants and students throughout lectures, particularly during exams. Establishing robust approaches and technologies is essential for identifying unjust, unethical, and illegal conduct in educational environments, particularly during academic courses and examinations. This study introduces a novel online proctoring system that employs deep learning to continuously supervise physical locations without requiring a human proctor. The system utilizes face detection and recognition algorithms to implement biometric procedures, including facial recognition. This work presents a novel approach for face recognition training, which involves an incremental training procedure. This eliminates the requirement for an additional step, resulting in reduced computation cost and time. In order to achieve high accuracy, the suggested model evaluated three distinct face detectors: HOG, MTCNN, and Yoloface. The assessment of the suggested model demonstrates that the HOG approach has surpassed the others. The proposed approach yields a significantly high accuracy rate.

MODNO: Multi-Operator learning with Distributed Neural Operators

The study of operator learning involves the utilization of neural networks to approximate operators. Traditionally, the focus has been on single-operator learning (SOL). However, recent advances have rapidly expanded this to include the approximation of multiple operators using foundation models equipped with millions or billions of trainable parameters, leading to the research of multi-operator learning (MOL). In this paper, we present a novel distributed training approach aimed at enabling a single neural operator with significantly fewer parameters to effectively tackle multi-operator learning challenges, all without incurring additional average costs. Our method is applicable to various neural operators, such as the Deep Operator Neural Networks (DON). The core idea is to independently learn the output basis functions for each operator using its dedicated data, while simultaneously centralizing the learning of the input function encoding shared by all operators using the entire dataset. Through a systematic study of five numerical examples, we compare the accuracy and cost of training a single neural operator for each operator independently versus training a MOL model using our proposed method. Our results demonstrate enhanced efficiency and satisfactory accuracy. Moreover, our approach illustrates that some operators with limited data can be more effectively constructed with the aid of data from analogous operators through MOL learning. This highlights another MOL’s potential to bolster operator learning.

Enhanced 3D dose prediction for hypofractionated SRS (gamma knife radiosurgery) in brain tumor using cascaded-deep-supervised convolutional neural network.

Gamma Knife radiosurgery (GKRS) is a well-established technique in radiation therapy (RT) for treating small-size brain tumors. It administers highly concentrated doses during each treatment fraction, with even minor dose errors posing a significant risk of causing severe damage to healthy tissues. It underscores the critical need for precise and meticulous precision in GKRS. However, the planning process for GKRS is complex and time-consuming, heavily reliant on the expertise of medical physicists. Incorporating deep learning approaches for GKRS dose prediction can reduce this dependency, improve planning efficiency and homogeneity, streamline clinical workflows, and reduce patient lagging times. Despite this, precise Gamma Knife plan dose distribution prediction using existing models remains a significant challenge. The complexity stems from the intricate nature of dose distributions, subtle contrasts in CT scans, and the interdependence of dosimetric metrics. To overcome these challenges, we have developed a "Cascaded-Deep-Supervised" Convolutional Neural Network (CDS-CNN) that employs a hybrid-weighted optimization scheme. Our innovative method incorporates multi-level deep supervision and a strategic sequential multi-network training approach. It enables the extraction of intra-slice and inter-slice features, leading to more realistic dose predictions with additional contextual information. CDS-CNN was trained and evaluated using data from 105 brain cancer patients who underwent GKRS treatment, with 85 cases used for training and 20 for testing. Quantitative assessments and statistical analyses demonstrated high consistency between the predicted dose distributions and the reference doses from the treatment planning system (TPS). The 3D overall gamma passing rates (GPRs) reached 97.15% ± 1.36% (3mm/3%, 10% threshold), surpassing the previous best performance by 2.53% using the 3D Dense U-Net model. When evaluated against more stringent criteria (2mm/3%, 10% threshold, and 1mm/3%, 10% threshold), the overall GPRs still achieved 96.53% ± 1.08% and 95.03% ± 1.18%. Furthermore, the average target coverage (TC) was 98.33% ± 1.16%, dose selectivity (DS) was 0.57 ± 0.10, gradient index (GI) was 2.69 ± 0.30, and homogeneity index (HI) was 1.79 ± 0.09. Compared to the 3D Dense U-Net, CDS-CNN predictions demonstrated a 3.5% improvement in TC, and CDS-CNN's dose prediction yielded better outcomes than the 3D Dense U-Net across all evaluation criteria. The experimental results demonstrated that the proposed CDS-CNN model outperformed other models in predicting GKRS dose distributions, with predictions closely matching the TPS doses.