A5054131772- Radiomics and Machine Learning in Medical Imaging
- Medical Image Segmentation Techniques
- Aortic aneurysm repair treatments
- Cardiac Valve Diseases and Treatments
- Medical Imaging Techniques and Applications
- Medical Imaging and Analysis
- Aortic Thrombus and Embolism
- Advanced X-ray and CT Imaging
- Cardiovascular Function and Risk Factors
- Advanced Radiotherapy Techniques
- Coronary Interventions and Diagnostics
- Lung Cancer Diagnosis and Treatment
- Cerebrovascular and Carotid Artery Diseases
- Industrial Vision Systems and Defect Detection
- Infective Endocarditis Diagnosis and Management
- AI in cancer detection
- Brain Tumor Detection and Classification
- Radiation Therapy and Dosimetry
- 3D Shape Modeling and Analysis
- Colorectal Cancer Surgical Treatments
- Gastric Cancer Management and Outcomes
- Esophageal Cancer Research and Treatment
- Digital Imaging for Blood Diseases
- Educational and Social Studies
- Higher Education Learning Practices
Cardiff University
2018-2021
Velindre Cancer Centre
2019-2021
University of Huddersfield
2021
Centro Cardiologico Monzino
2016-2020
Politecnico di Milano
2013-2016
Swiss Institute for Regenerative Medicine
2016
Università della Svizzera italiana
2014-2015
University of Chicago Medical Center
2013
The aim of this work was to investigate radiomic analysis contrast and non-contrast enhanced planning CT images oesophageal cancer (OC) patients in terms stability, dimensionality agent dependency. prognostic significance CT-based features also evaluated. Different 2D 3D were extracted from 213 the multi-centre SCOPE1 randomised controlled trial (RCT) OC. Feature stability evaluated by randomly dividing into three groups identifying textures with similar distributions among a Kruskal-Wallis...
To propose a nearly automated left ventricular (LV) three-dimensional (3D) surface segmentation procedure, based on active shape modelling (ASM) and built database of 3D echocardiographic (3DE) LV surfaces, for cardiac magnetic resonance (CMR) images, to test its accuracy volumes computation compared with ‘gold standard’ manual tracings discs-summation method. The ASM was created segmented surfaces (4D analysis, Tomtec) from 3DE datasets 205 patients. Then, it applied the imaging short-axis...
Accurate classification of haematological cells is critical for diagnosing blood disorders, but presents significant challenges machine automation owing to the complexity cell morphology, heterogeneities biological, pathological, and imaging characteristics, imbalance type frequencies. We introduce CytoDiffusion, a diffusion-based classifier that effectively models combining accurate with robust anomaly detection, resistance distributional shifts, interpretability, data efficiency,...
Target volume delineation (TVD) has been identified as a weakness in the accuracy of radiotherapy, both within and outside clinical trials due to intra/interobserver variations affecting TVD quality. Sources such poor compliance or protocol violation may have adverse effect on treatment outcomes. In this paper, we present describe FIELDRT software developed for ARENA project improve quality through qualitative quantitative feedbacks individual personalized summary trainee"s performance.For...
Statistical shape models (SSMs) represent a powerful tool used in patient-specific modeling to segment medical images because they incorporate a-priori knowledge that guide the model during deformation. Our aim was evaluate segmentation accuracy terms of left ventricular (LV) volumes obtained using four different SSMs versus manual gold standard tracing on cardiac magnetic resonance (CMR) images. A database 3D echocardiographic (3DE) LV surfaces 435 patients generate SSMs, based phase...
Statistical shape modelling (SSM) approaches have been proposed as a powerful tool to segment the left ventricle in cardiac magnetic resonance (CMR) images.Our aim was extend this method RV cavity CMR images and validate it compared conventional gold-standard (GS) manual tracing.A SSM of built using database 4347 intrinsically 3D surfaces, extracted from transthoracic echocardiographic (3DE) 219 retrospective patients.The then scaled deformed on base some features extracted, with different...
The results indicate both a good accuracy and generality of the trained model.ML model seems to be able predict when plan solution is deliverable during planning processes, reducing number DQA failures, re-optimizations due failing plans being rejected physics check because overmodulation issue.