Results 121 to 130 of about 118,357 (248)

An Improved ResNet Based on the Adjustable Shortcut Connections

IEEE Access, 2018
ResNet can achieve deeper network and higher performance, but there is no good explanation for how identity shortcut connections solve the gradient fading problems. Moreover, it is not reasonable to adopt identity mapping for all layer parameters.
Baoqi Li, Yuyao He
doaj   +1 more source

IAR‐Net: Tabular Deep Learning Model for Interventionalist's Action Recognition

Advanced Intelligent Systems, EarlyView.
This study presents IAR‐Net, a deep‐learning framework for catheterization action recognition. To ensure optimality, this study quantifies interoperator similarities and differences using statistical tests, evaluates the distribution fidelity of synthetic data produced by six generative models, and benchmarks multiple deep‐learning models.
Toluwanimi Akinyemi, Olatunji Omisore, Wenjing Du, Wenke Duan, Chen Bailang, Liu Kun, Lei Wang, Minxin Wei   +7 more
wiley   +1 more source

Autonomous Recognition of Retained Secretions in Central‐Airway Based on Deep Learning for Adult Patients Receiving Invasive Mechanical Ventilation

Advanced Intelligent Systems, EarlyView.
This work presents a deep learning model to autonomously recognize and classify the secretion retention into three levels for patients receiving invasive mechanical ventilation, achieving 89.08% accuracy. This model can be implemented to ventilators by edge computing, whose feasibility is approved.
Shuai Wang, Zhuoran Xiao, Meng Wang, Jianchao Li, Chunjing Tao, Jiangzhen Guo, Yubo Fan   +6 more
wiley   +1 more source

Loss‐Based Ensemble Generative Adversarial Network Model for Enhancing the Sperm Morphology Classification

Advanced Intelligent Systems, EarlyView.
A loss‐based ensemble generative adversarial network (GAN) framework is proposed to address mode collapse in sperm morphology classification. By integrating spatial augmentation and multiple GAN models, the study enhances synthetic data quality. The Shifted Window Transformer achieves 95.37% accuracy on the HuSHeM dataset, outperforming previous ...
Berke Cansiz, Hamza Osman Ilhan, Gorkem Serbes   +2 more
wiley   +1 more source

Deep learning-based dual optimization framework for accurate thyroid disease diagnosis using CNN architectures

Mehran University Research Journal of Engineering and Technology
Thyroid diseases, including hypothyroidism, hyperthyroidism, thyroid nodules, thyroiditis, and thyroid cancer, are among the most prevalent endocrine disorders, posing significant health risks, which need to be diagnosed and treated promptly. Traditional
Zeeshan Ali Haider, Nasser A Alsadhan, Fida Muhammad Khan, Waleed Al-Azzawi, Inam Ullah Khan, Inam Ullah   +5 more
doaj   +1 more source

Impact of Biomimetic Pinna Shape Variation on Clutter Echoes: A Machine Learning Approach

Advanced Intelligent Systems, EarlyView.
Bats with dynamic ear structures navigate dense, echo‐rich environments, yet the echoes they receive are highly random. This study shows that machine learning can reliably detect structural signatures in these seemingly chaotic biosonar signals. The results open new directions for biologically inspired sensing, where time‐varying receiver shapes ...
Ibrahim Eshera, Sanmeel Lagad, Rolf Müller   +2 more
wiley   +1 more source

Multitarget Recognition of Flower Images Based on Lightweight Deep Neural Network and Transfer Learning

Advanced Intelligent Systems, EarlyView.
This article proposes a lightweight YOLOv4‐based detection model using MobileNetV3 or CSPDarknet53_tiny, achieving 30+ FPS and higher mAP. It also presents a ShuffleNet‐based classification model with transfer learning and GAN‐augmented images, improving generalization and accuracy.
Qingyang Liu, Yanrong Hu, Hongjiu Liu
wiley   +1 more source

From Droplet to Diagnosis: Spatio‐Temporal Pattern Recognition in Drying Biofluids

Advanced Intelligent Systems, EarlyView.
This article integrates machine learning (ML) with the spatio‐temporal evolution of biofluid droplets to reveal how drying and self‐assembly encode distinctive compositional fingerprints. By leveraging textural features and interpretable ML, it achieves robust classification of blood abnormalities with over 95% accuracy.
Anusuya Pal, Amalesh Gope, Miho Yanagisawa   +2 more
wiley   +1 more source

Impact of using various x-ray dataset in detecting tuberculosis based on deep learning

Радіоелектронні і комп'ютерні системи
The subject matter is that the characteristics of tuberculosis are difficult to study visually. Therefore, a computer-aided system based on deep learning can be applied to X-ray image recognition.
Muhammad Irhamsyah, Qurrata A’yuni, Khairun Saddami, Nasaruddin Nasaruddin, Khairul Munadi, Fitri Arnia   +5 more
doaj   +1 more source

Cross‐Modal Characterization of Thin‐Film MoS2 Using Generative Models

Advanced Intelligent Systems, EarlyView.
Cross‐modal learning is evaluated using atomic force microscopy (AFM), Raman spectroscopy, and photoluminescence spectroscopy (PL) through unsupervised learning, regression, and autoencoder models. Autoencoder models are used to generate spectroscopy data from the microscopy images.
Isaiah A. Moses, Chen Chen, Joan M. Redwing, Wesley F. Reinhart   +3 more
wiley   +1 more source