STIRNet: A Spatio-Temporal Network for Air Formation Targets Intention Recognition DOI Creative Commons
Chenhao Zhang, Yan Zhou, Hongquan Li

и другие.

IEEE Access, Год журнала: 2024, Номер 12, С. 44998 - 45010

Опубликована: Янв. 1, 2024

Air formation combat is a common style of air combat, which demonstrates high degree flexibility and strategic value in complex battlefield environments.The activity state the result intertwining time domain domain, requires accurate execution tactical processes axis skillful deployment forces three-dimensional space.Therefore, target intention recognition challenging task that an in-depth understanding dynamically changing behavioral patterns formation.To address this problem, paper proposes STIRNet (Spatio-Temporal Intention Recognition Network) model, abstracts as spatial graph structure composed vehicle nodes combines its temporal data evolving over time.The model autonomously adjusts attention to different moments locations through spatio-temporal mechanism, focusing on important features are crucial for recognizing formation; simultaneously captures integrates feature information both dimensions spatiotemporal convolutional operation, effectively solves deficiencies traditional methods dealing with dependency relationships.The experimental results show proposed improves accuracy targets, great command decision-making situation assessment.

Язык: Английский

Enhancing Medical Image Denoising with Innovative Teacher–Student Model-Based Approaches for Precision Diagnostics DOI Creative Commons
Shakhnoza Muksimova, Sabina Umirzakova,

Sevara Mardieva

и другие.

Sensors, Год журнала: 2023, Номер 23(23), С. 9502 - 9502

Опубликована: Ноя. 29, 2023

The realm of medical imaging is a critical frontier in precision diagnostics, where the clarity image paramount. Despite advancements technology, noise remains pervasive challenge that can obscure crucial details and impede accurate diagnoses. Addressing this, we introduce novel teacher–student network model leverages potency our bespoke NoiseContextNet Block to discern mitigate with unprecedented precision. This innovation coupled an iterative pruning technique aimed at refining for heightened computational efficiency without compromising fidelity denoising. We substantiate superiority effectiveness approach through comprehensive suite experiments, showcasing significant qualitative enhancements across multitude modalities. visual results from vast array tests firmly establish method’s dominance producing clearer, more reliable images diagnostic purposes, thereby setting new benchmark

Язык: Английский

Процитировано

25
CFATransUnet: Channel-wise cross fusion attention and transformer for 2D medical image segmentation DOI
Cheng Wang, Le Wang,

Nuoqi Wang

и другие.

Computers in Biology and Medicine, Год журнала: 2023, Номер 168, С. 107803 - 107803

Опубликована: Дек. 4, 2023

Язык: Английский

Процитировано

16
Facial Recognition Algorithms: A Systematic Literature Review DOI Creative Commons

N Fadel

Journal of Imaging, Год журнала: 2025, Номер 11(2), С. 58 - 58

Опубликована: Фев. 13, 2025

This systematic literature review aims to understand new developments and challenges in facial recognition technology. will provide an understanding of the system principles, performance metrics, applications technology various fields such as health, society, security from academic publications, conferences, industry news. A comprehensive approach was adopted technologies. It emphasizes most important techniques algorithm development, examines explores their fields. The mainly recent development deep learning techniques, especially CNNs, which greatly improved accuracy efficiency systems. findings reveal that there has been a noticeable evolution technology, with current use techniques. Nevertheless, it highlights challenges, including privacy concerns, ethical dilemmas, biases These factors highlight necessity using regulated manner. In conclusion, paper proposes several future research directions establish reliability systems reduce while building user confidence. considerations are key responsibly advancing by ensuring practices safeguarding privacy.

Язык: Английский

Процитировано

0
YOLO-TARC: YOLOv10 with Token Attention and Residual Convolution for Small Void Detection in Root Canal X-Ray Images DOI Creative Commons
Pan Yin,

Zhenpeng Zhang,

Xueyang Zhang

и другие.

Sensors, Год журнала: 2025, Номер 25(10), С. 3036 - 3036

Опубликована: Май 12, 2025

The detection of small voids or defects in X-ray images tooth root canals still faces challenges. To address the issue, this paper proposes an improved YOLOv10 that combines Token Attention with Residual Convolution (ResConv), termed YOLO-TARC. overcome limitations existing deep learning models effectively retaining key features objects and their insufficient focusing capabilities, we introduce three improvements. First, ResConv is designed to ensure transmission discriminative during feature propagation, leveraging ability residual connections transmit information from one layer next. Second, tackle issue weak capabilities on targets, a module introduced before third object head. By tokenizing maps enhancing local focusing, it enables model pay closer attention targets. Additionally, optimize training process, bounding box loss function adopted achieve faster more accurate predictions. YOLO-TARC simultaneously enhances retain detailed targets improves thereby increasing accuracy. Experimental results private canal image dataset demonstrate outperforms other state-of-the-art models, achieving 7.5% improvement 80.8% mAP50 6.2% increase 80.0% Recall. can contribute efficient objective postoperative evaluation treatments.

Язык: Английский

Процитировано

0
MRMS-CNNFormer: A Novel Framework for Predicting the Biochemical Recurrence of Prostate Cancer on Multi-Sequence MRI DOI Creative Commons
Tao Lian, Mengting Zhou, Yi Ming Shao

и другие.

Bioengineering, Год журнала: 2025, Номер 12(5), С. 538 - 538

Опубликована: Май 16, 2025

Accurate preoperative prediction of biochemical recurrence (BCR) in prostate cancer (PCa) is essential for treatment optimization, and demands an explicit focus on tumor microenvironment (TME). To address this, we developed MRMS-CNNFormer, innovative framework integrating 2D multi-region (intratumoral, peritumoral, periprostatic) multi-sequence magnetic resonance imaging (MRI) images (T2-weighted with fat suppression (T2WI-FS) diffusion-weighted (DWI)) clinical characteristics. The utilizes a CNN-based encoder feature extraction, followed by transformer-based multi-modal integration, ultimately employs fully connected (FC) layer final BCR prediction. In this multi-center study (46 BCR-positive cases, 186 BCR-negative cases), patients from centers A B were allocated to training (n = 146) validation 36) sets, while center C 50) formed the external test set. MRI-based model demonstrated superior performance (AUC, 0.825; 95% CI, 0.808–0.852) compared single-region models. integration data further enhanced model’s predictive capability (AUC 0.835; 0.818–0.869), significantly outperforming alone 0.612; 0.574–0.646). MRMS-CNNFormer provides robust, non-invasive approach prediction, offering valuable insights personalized planning decision making PCa management.

Язык: Английский

Процитировано

0
An Improved Rotating Box Detection Model for Litchi Detection in Natural Dense Orchards DOI Creative Commons
Bin Li, Huazhong Lu, Xinyu Wei

и другие.

Agronomy, Год журнала: 2023, Номер 14(1), С. 95 - 95

Опубликована: Дек. 30, 2023

Accurate litchi identification is of great significance for orchard yield estimations. Litchi in natural scenes have large differences scale and are occluded by leaves, reducing the accuracy detection models. Adopting traditional horizontal bounding boxes will introduce a amount background overlap with adjacent frames, resulting reduced accuracy. Therefore, this study innovatively introduces use rotation box model to explore its capabilities scenarios occlusion small targets. First, dataset on constructed. Secondly, three improvement modules based YOLOv8n proposed: transformer module introduced after C2f eighth layer backbone network, an ECA attention added neck network improve feature extraction 160 × head enhance target detection. The test results show that, compared model, proposed improves precision rate, recall mAP 11.7%, 5.4%, 7.3%, respectively. In addition, four state-of-the-art mainstream networks, namely, MobileNetv3-small, MobileNetv3-large, ShuffleNetv2, GhostNet, studied comparison performance model. article exhibits better dataset, precision, recall, reaching 84.6%, 68.6%, 79.4%, This research can provide reference estimations complex environments.

Язык: Английский

Процитировано

2
STIRNet: A Spatio-Temporal Network for Air Formation Targets Intention Recognition DOI Creative Commons
Chenhao Zhang, Yan Zhou, Hongquan Li

и другие.

IEEE Access, Год журнала: 2024, Номер 12, С. 44998 - 45010

Опубликована: Янв. 1, 2024

Air formation combat is a common style of air combat, which demonstrates high degree flexibility and strategic value in complex battlefield environments.The activity state the result intertwining time domain domain, requires accurate execution tactical processes axis skillful deployment forces three-dimensional space.Therefore, target intention recognition challenging task that an in-depth understanding dynamically changing behavioral patterns formation.To address this problem, paper proposes STIRNet (Spatio-Temporal Intention Recognition Network) model, abstracts as spatial graph structure composed vehicle nodes combines its temporal data evolving over time.The model autonomously adjusts attention to different moments locations through spatio-temporal mechanism, focusing on important features are crucial for recognizing formation; simultaneously captures integrates feature information both dimensions spatiotemporal convolutional operation, effectively solves deficiencies traditional methods dealing with dependency relationships.The experimental results show proposed improves accuracy targets, great command decision-making situation assessment.

Язык: Английский

Процитировано

0