A5054804413- Medical Image Segmentation Techniques
- Advanced Neural Network Applications
- Brain Tumor Detection and Classification
- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Cancer Immunotherapy and Biomarkers
- Digital Imaging for Blood Diseases
- Cancer Genomics and Diagnostics
- Head and Neck Cancer Studies
- Brain Metastases and Treatment
- Retinal Imaging and Analysis
- Lung Cancer Treatments and Mutations
- Physical Unclonable Functions (PUFs) and Hardware Security
- Advanced Image and Video Retrieval Techniques
- Optical Coherence Tomography Applications
- Cell Image Analysis Techniques
- Security and Verification in Computing
- Radiation Effects in Electronics
- Explainable Artificial Intelligence (XAI)
- Acute Ischemic Stroke Management
- Context-Aware Activity Recognition Systems
- Glaucoma and retinal disorders
- Domain Adaptation and Few-Shot Learning
- Kidney Stones and Urolithiasis Treatments
- Ferroptosis and cancer prognosis
Lutron Electronics (United States)
2022-2023
University of Minho
2014-2022
Iscte – Instituto Universitário de Lisboa
2019
Hospital Universitario Ramón y Cajal
2014
Medical University of Silesia
2014
Ulyanovsk State University
2014
GlaxoSmithKline (Brazil)
2003
Among brain tumors, gliomas are the most common and aggressive, leading to a very short life expectancy in their highest grade. Thus, treatment planning is key stage improve quality of oncological patients. Magnetic resonance imaging (MRI) widely used technique assess these but large amount data produced by MRI prevents manual segmentation reasonable time, limiting use precise quantitative measurements clinical practice. So, automatic reliable methods required; however, spatial structural...
Many methods have been proposed for tissue segmentation in brain MRI scans. The multitude of complicates the choice one method above others. We therefore established MRBrainS online evaluation framework evaluating (semi)automatic algorithms that segment gray matter (GM), white (WM), and cerebrospinal fluid (CSF) on 3T scans elderly subjects (65–80 y). Participants apply their to provided data, after which results are evaluated ranked. Full manual segmentations GM, WM, CSF available all used...
Biomarkers on the basis of tumor-infiltrating lymphocytes (TIL) are potentially valuable in predicting effectiveness immune checkpoint inhibitors (ICI). However, clinical application remains challenging because methodologic limitations and laborious process involved spatial analysis TIL distribution whole-slide images (WSI).We have developed an artificial intelligence (AI)-powered WSI analyzer tumor microenvironment that can define three phenotypes (IPs): inflamed, immune-excluded,...
Computational pathology can lead to saving human lives, but models are annotation hungry and images notoriously expensive annotate. Self-supervised learning (SSL) has shown be an effective method for utilizing unlabeled data, its application could greatly benefit downstream tasks. Yet, there no principled studies that compare SSL methods discuss how adapt them pathology. To address this need, we execute the largest-scale study of pre-training on image date. Our is conducted using 4...
Gliomas are among the most common and aggressive brain tumours. Segmentation of these tumours is important for surgery treatment planning, but also follow-up evaluations. However, it a difficult task, given that its size locations variable, delineation all tumour tissue not trivial, even with different modalities Magnetic Resonance Imaging (MRI). We propose discriminative fully automatic method segmentation gliomas, using appearance- context-based features to feed an Extremely Randomized...
Cell detection is a fundamental task in computational pathology that can be used for extracting high-level medical information from whole-slide images. For accurate cell detection, pathologists often zoom out to understand the tissue-level structures and classify cells based on their morphology surrounding context. However, there lack of efforts reflect such behaviors by models, mainly due datasets containing both tissue annotations with overlapping regions. To overcome this limitation, we...
Abstract Background Accurate classification of breast cancer molecular subtypes is crucial in determining treatment strategies and predicting clinical outcomes. This largely depends on the assessment human epidermal growth factor receptor 2 (HER2), estrogen (ER), progesterone (PR) status. However, variability interpretation among pathologists pose challenges to accuracy this classification. study evaluates role artificial intelligence (AI) enhancing consistency these evaluations. Methods...
Abstract Tumor-infiltrating lymphocytes (TILs) have been recognized as key players in the tumor microenvironment of breast cancer, but substantial interobserver variability among pathologists has impeded its utility a biomarker. We developed deep learning (DL)-based TIL analyzer to evaluate stromal TILs (sTILs) cancer. Three evaluated 402 whole slide images cancer and interpreted sTIL scores. A standalone performance DL model was 210 cases (52.2%) exhibiting score differences less than 10...
Evaluation of PD-L1 tumor proportion score (TPS) by pathologists has been very impactful but is limited factors such as intraobserver/interobserver bias and intratumor heterogeneity. We developed an artificial intelligence (AI)-powered analyzer to assess TPS for the prediction immune checkpoint inhibitor (ICI) response in advanced non-small cell lung cancer (NSCLC).
Fully Convolutional Networks have been achieving remarkable results in image semantic segmentation, while being efficient. Such efficiency from the capability of segmenting several voxels a single forward pass. So, there is direct spatial correspondence between unit feature map and voxel same location. In convolutional layer, kernel spans over all channels extracts information them. We observe that linear recombination maps by increasing number followed compression may enhance their...
The retinal vascular condition is a trustworthy biomarker of several ophthalmologic and cardiovascular diseases, so automatic vessel segmentation crucial step to diagnose monitor these problems. Deep Learning models have recently revolutionized the state-of-the-art in fields, since they can learn features with multiple levels abstraction from data itself. However, methods easily fall into overfitting, huge number parameters must be learned. Having bigger datasets may act as regularization...
Magnetic Resonance Imaging is the preferred imaging modality for assessing brain tumors, and segmentation necessary diagnosis treatment planning. Thus, robust automatic methods are required. Machine learning proposals where model learned from data quite successful. Hierarchical approaches firstly segment whole tumor, followed by intra-tumor tissue identification. However, results comparing it with single stages needed, as state of art also achieved all-at-once strategies. Currently, fully...