Results 121 to 130 of about 3,449,855 (345)
This study presents Hi‐LabSpermTracking, a long‐duration, expert‐annotated sperm motility dataset for detection and tracking. It evaluates You Only Look Once version 8, Real‐Time Detection Transformer, and Simple Online and Realtime Tracking with a Deep Association Metric in three scenarios. The mean ensemble method improves tracking performance. Sperm
Abdulsamet Aktas+5 more
wiley +1 more source
Lorna Gamberini is a Speech and Language Therapist with the Morecambe Bay Primary Care Trust. In this article she describes the possible links between dysphonia and reflux symptoms. It has been 'received wisdom' in her profession that there is an aetiological link, but this is a matter of debate with the medical profession.
openaire +2 more sources
Improving Long‐Term Glucose Prediction Accuracy with Uncertainty‐Estimated ProbSparse‐Transformer
Wearable devices collect blood glucose and other physiological data, which serve as inputs to the prediction model. After data embedding, a structure utilizing ProbSparse self‐attention and a one‐step generative head within a Transformer‐based model is introduced, which is concurrently designed for deployment on edge devices, enabling real‐time ...
Wei Huang+5 more
wiley +1 more source
Voice Disorder: case definition and prevalence in teachers [PDF]
Renata Jardim+2 more
openalex +1 more source
A novel medical image segmentation framework that integrates an enhanced Particle Swarm Optimization (EE‐PSO) into DeepLabV3+ to optimize feature selection. By dynamically identifying key channels in the atrous spatial pyramid pooling module, the method improves segmentation performance, achieving mIoU gains of 2.7% on Alzheimer's and 2.8% on brain ...
Alireza Norouziazad+8 more
wiley +1 more source
Screening of speech and voice disorders among school children in Itabashi-ku, Tokyo.
Yoshiaki YOKOMI+12 more
openalex +2 more sources
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+6 more
wiley +1 more source
Development of a Femininity Estimator using Speaker Recognition Techniques for Voice Therapy of Gender Identity Disorder Clients [PDF]
Nobuaki Minematsu+6 more
openalex +1 more source