A5080470110- AI in cancer detection
- Digital Radiography and Breast Imaging
- Image and Signal Denoising Methods
- Radiomics and Machine Learning in Medical Imaging
- Breast Cancer Treatment Studies
- Medical Imaging Techniques and Applications
- Infrared Thermography in Medicine
- 3D Surveying and Cultural Heritage
- Advanced X-ray and CT Imaging
- Digital Media Forensic Detection
- Cancer Genomics and Diagnostics
- Breast Lesions and Carcinomas
- Industrial Vision Systems and Defect Detection
Universidade de São Paulo
2019-2024
Lund University
2023
University of Groningen
2019
Steadily increasing use of computational/virtual phantoms in medical physics has motivated expanding development new simulation methods and data representations for modelling human anatomy. This emphasized the need increased realism, user control, availability. In breast cancer research, virtual have gained an important role evaluating optimizing imaging systems. For this paper, we developed algorithm to model abnormalities based on fractal Perlin noise. We demonstrate characterize extension...
Early detection of breast cancer can increase treatment efficiency. Architectural Distortion (AD) is a very subtle contraction the tissue and may represent earliest sign cancer. Since it likely to be unnoticed by radiologists, several approaches have been proposed over years but none using deep learning techniques. To train Convolutional Neural Network (CNN), which neural architecture, necessary huge amount data. overcome this problem, paper proposes data augmentation approach applied...
Deep learning models have reached superior results in various fields of application, but many cases at a high cost processing or large amount data available. In most them, specially the medical field, scarcity training limits performance these models. Among strategies to overcome lack data, there is augmentation, transfer and fine-tuning. this work we compared different approaches train deep convolutional neural network (CNN) automatically detect architectural distortion (AD) digital...
In previous work, we investigated the application of normalized anisotropic quality index (NAQI) as an image metric for digital mammography. The initial assessment showed that NAQI depends not only on radiation dose, but also varies based features such breast anatomy. this these dependencies are analyzed by assessing contribution a range values. generalized matrix learning vector quantization (GMLVQ) was used to evaluate feature relevance and rank imaging parameters affect NAQI. GMLVQ uses...
Steadily increasing use of computer-generated models sets high demands on the realistic representation breast tissue and abnormalities. Previously, we demonstrated Perlin noise to simulate lesions. Now, expand previous model by simulating heterogeneous lesion composition. We demonstrate a new approach 3D soft lesions, with additional benefits in context virtual clinical trials. Three simulation methods have been developed: Method I represents homogeneous made up glandular (our model). II...
The combination of digital breast tomosynthesis (DBT) with other imaging modalities has been investigated in order to improve the detection and diagnosis cancer. Mechanical Imaging (MI) measures stress over surface compressed breast, using a pressure sensor, during radiographic examination its response shown correlation presence malignant lesions. Thus, DBT MI (DBTMI) potential reduce false positive results cancer screening. However, compared conventional exam, sensor mammographic image...
The angular range and number of projections are parameters that directly influence the image quality visibility lesions in digital breast tomosynthesis (DBT). medical field is taking advantage increasing performance machine learning algorithms with use complex data-driven models, known as deep (DL) networks. DL has also been highlighted tasks video frame interpolation (VFI) for synthesis new images order to increase rate per second. In present work, we a residual refinement network (RRIN)...