Browsing by Author "Salas, Rodrigo"

Now showing 1 - 4 of 4
  • No Thumbnail Available
    Item
    Benchmarking YOLO Models for Intracranial Hemorrhage Detection Using Varied CT Data Sources
    (IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 2024) Tapia, Gonzalo; Allende-Cid, Hector; Chabert, Steren; Mery Quiroz Domingo Arturo; Salas, Rodrigo
    Intracranial hemorrhages (ICH) are a significant challenge in emergency medicine due to the critical nature of a timely and accurate diagnosis. This study evaluates the performance of six versions of the You Only Look Once (YOLO) object detection model, from YOLOv5 to YOLOv10, in detecting ICH using computed tomography (CT) scans. The primary focus is understanding the advancements in YOLO architectures over time and their impact on detection accuracy and inference speed. The study used the Brain Hemorrhage Extended Dataset (BHX), comprising 491 CT scans with annotations for six types of hemorrhages: epidural, subdural, subarachnoid, intraparenchymal, intraventricular, and chronic hemorrhage, and introduces a new data set obtained from a major hospital in Chile. The models were trained using a combination of single-class and multi-class approaches to address class imbalance and were evaluated based on precision, recall, F1 score, and mean average precision (mAP). The models were evaluated in three distinct contexts: 1) a biased scenario where images of the same individual could appear in both training and testing sets, 2) a cross-validation setup ensuring the independence of images by separating the sets based on subjects, and 3) an external validation using one dataset for training and the Chilean dataset for testing, maintaining full independence between training and evaluation. The findings indicate that YOLOv8 and YOLOv10 demonstrate superior detection accuracy and inference efficiency performance, respectively, compared to previous versions. In particular, with image independence, YOLOv8 reached the highest average mAP for all classes, with a score of 0.4. This comparative analysis provides information on the effectiveness of architectural advances in YOLO models for medical applications and suggests directions for future improvements in ICH detection.
  • No Thumbnail Available
    Item
    Glycaemia dynamics in gestational diabetes mellitus
    (2022) Valero, Paola; Salas, Rodrigo; Pardo, Fabian; Cornejo, Marcelo; Fuentes, Gonzalo; Vega, Sofia; Grismaldo, Adriana; Hillebrands, Jan-Luuk; van der Beek, Eline M.; van Goor, Harry; Sobrevia, Luis
    Pregnant women may develop gestational diabetes mellitus (GDM), a disease of pregnancy characterised by maternal and fetal hyperglycaemia with hazardous consequences to the mother, the fetus, and the newborn. Maternal hyperglycaemia in GDM results in fetoplacental endothelial dysfunction. GDM-harmful effects result from chronic and short periods of hyperglycaemia. Thus, it is determinant to keep glycaemia within physiological ranges avoiding short but repetitive periods of hyper or hypoglycaemia. The variation of glycaemia over time is defined as 'glycaemia dynamics'. The latter concept regards with a variety of mechanisms and environmental conditions leading to blood glucose handling. In this review we summarized the different metrics for glycaemia dynamics derived from quantitative, plane distribution, amplitude, score values, variability estimation, and time series analysis. The potential application of the derived metrics from self-monitoring of blood glucose (SMBG) and continuous glucose monitoring (CGM) in the potential alterations of pregnancy outcome in GDM are discussed.
  • Thumbnail Image
    Item
    Identification of hemodynamic biomarkers for bicuspid aortic valve induced aortic dilation using machine learning
    (2022) Franco, Pamela; Sotelo, Julio; Guala, Andrea; Dux-Santoy, Lydia; Evangelista, Arturo; Rodriguez-Palomares, Jose; Mery Quiroz, Domingo Arturo; Salas, Rodrigo; Uribe, Sergio
  • Thumbnail Image
    Item
    Quantitative description of the morphology and ossification center in the axial skeleton of 20-week gestation formalin-fixed human fetuses using magnetic resonance images
    (WILEY, 2012) Chabert, Steren; Villalobos, Manuel; Ulloa, Patricia; Salas, Rodrigo; Tejos, Cristian; San Martin, Sebastian; Pereda, Jaime
    Objectives Human tissues are usually studied using a series of two-dimensional visualizations of in vivo or cutout specimens. However, there is no precise anatomical description of some of the processes of human fetal development. The purpose of our study is to develop a quantitative description of the normal axial skeleton by means of high-resolution three-dimensional magnetic resonance (MR) images, collected from six normal 20-week-old human fetuses fixed in formaldehyde.