A5047066399- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Generative Adversarial Networks and Image Synthesis
- Advanced Vision and Imaging
- Advanced Image Processing Techniques
- Colorectal Cancer Screening and Detection
- MRI in cancer diagnosis
- Image Enhancement Techniques
- Computer Graphics and Visualization Techniques
- 3D Shape Modeling and Analysis
- Advanced Image and Video Retrieval Techniques
- Digital Media Forensic Detection
- Computational Geometry and Mesh Generation
- Topic Modeling
- Advanced Image Fusion Techniques
- Advanced Neural Network Applications
- Music and Audio Processing
- Data Visualization and Analytics
- Actinomycetales infections and treatment
- Oral and Maxillofacial Pathology
- Autonomous Vehicle Technology and Safety
- Model Reduction and Neural Networks
- Infectious Diseases and Mycology
- Artificial Intelligence in Healthcare and Education
- 3D Surveying and Cultural Heritage
Karolinska Institutet
2023-2025
Linköping University
2014-2022
Artificial Intelligence (AI) research in breast cancer Magnetic Resonance Imaging (MRI) faces challenges due to limited expert-labeled segmentations. To address this, we present a multicenter dataset of 1506 pre-treatment T1-weighted dynamic contrast-enhanced MRI cases, including expert annotations primary tumors and non-mass-enhanced regions. The integrates imaging data from four collections Cancer Archive (TCIA), where only 163 cases with segmentations were initially available. facilitate...
Abstract Image synthesis designed for machine learning applications provides the means to efficiently generate large quantities of training data while controlling generation process provide best distribution and content variety. With demands deep applications, synthetic have potential becoming a vital component in pipeline. Over last decade, wide variety methods has been demonstrated. The future development calls bring these together comparison categorization. This survey comprehensive list...
Despite its benefits for tumour detection and treatment, the administration of contrast agents in dynamic contrast-enhanced MRI (DCE-MRI) is associated with a range issues, including their invasiveness, bioaccumulation, risk nephrogenic systemic fibrosis. This study explores feasibility producing synthetic enhancements by translating pre-contrast T1-weighted fat-saturated breast to corresponding first DCE-MRI sequence leveraging capabilities generative adversarial network (GAN)....
We present an overview and evaluation of a new, systematic approach for generation highly realistic, annotated synthetic data training deep neural networks in computer vision tasks. The main contribution is procedural world modeling enabling high variability coupled with physically accurate image synthesis, departure from the hand-modeled virtual worlds approximate synthesis methods used real-time applications. benefits our include flexible, scalable implicit wide coverage classes features,...
Single-image high dynamic range (SI-HDR) reconstruction has recently emerged as a problem well-suited for deep learning methods. Each successive technique demonstrates an improvement over existing methods by reporting higher image quality scores. This paper, however, highlights that such improvements in objective metrics do not necessarily translate to visually superior images. The first is the use of disparate evaluation conditions terms data and metric parameters, calling standardized...
PurposeMultiple vendors are currently offering artificial intelligence (AI) computer-aided systems for triage detection, diagnosis, and risk prediction of breast cancer based on screening mammography. There is an imminent need to establish validation platforms that enable fair transparent testing these against external data.ApproachWe developed imaging (VAI-B), a platform independent AI algorithms in imaging. The hybrid solution, with one part implemented the cloud another on-premises...
Current research in breast cancer Magnetic Resonance Imaging (MRI), especially with Artificial Intelligence (AI), faces challenges due to the lack of expert segmentations. To address this, we introduce MAMA-MIA dataset, comprising 1506 multi-center dynamic contrast-enhanced MRI cases segmentations primary tumors and non-mass enhancement areas. These were sourced from four publicly available collections The Cancer Archive (TCIA). Initially, trained a deep learning model automatically segment...
Insufficient training data is a major bottleneck for most deep learning practices, not least in medical imaging where difficult to collect and publicly available datasets are scarce due ethics privacy. This work investigates the use of synthetic images, created by generative adversarial networks (GANs), as only source data. We demonstrate that this application, it great importance make multiple GANs improve diversity generated data, i.e. sufficiently cover distribution. While single GAN can...
Data driven reflectance models using BRDF data measured from real materials, e.g. [Matusik et al. 2003], are becoming increasingly popular in product visualization, digital design and other applications by the need for predictable rendering highly realistic results. Although recent analytic, parametric BRDFs provide good approximations many some effects still not captured well [Löw 2012]. Thus, it is hard to accurately model materials analytic models, even if parameters fitted data. In...
Digitization of histopathology slides has led to several advances, from easy data sharing and collaborations the development digital diagnostic tools. Deep learning (DL) methods for classification detection have shown great potential, but often require large amounts training that are hard collect, annotate. For many cancer types, scarceness creates barriers DL models. One such scenario relates detecting tumor metastasis in lymph node tissue, where low ratio non-tumor cells makes task...
Contrast agents in dynamic contrast enhanced magnetic resonance imaging allow to localize tumors and observe their kinetics, which is essential for cancer characterization respective treatment decision-making. However, agent administration not only associated with adverse health risks, but also restricted patients during pregnancy, those kidney malfunction, or other reactions. With uptake as key biomarker lesion malignancy, recurrence risk, response, it becomes pivotal reduce the dependency...
Medical image segmentation has improved with deep-learning methods, especially for tumor segmentation. However, variability in shapes, sizes, and enhancement remains a challenge. Breast MRI adds further uncertainty due to anatomical differences. Informing clinicians about result reliability using model improve predictions are essential. We study Monte-Carlo Dropout generating multiple finding consensus Our approach reduces false positives per-pixel improves metrics. In addition, we the...
This paper presents a method for virtual contrast enhancement in breast MRI, offering promising non-invasive alternative to traditional agent-based DCE-MRI acquisition. Using conditional generative adversarial network, we predict images, including jointly-generated sequences of multiple corresponding timepoints, from non-contrast-enhanced MRIs, enabling tumor localization and characterization without the associated health risks. Furthermore, qualitatively quantitatively evaluate synthetic...
This paper presents an evaluation of how data augmentation and inter-class transformations can be used to synthesize training in low-data scenarios for single-image weather classification. In such scenarios, augmentations is a critical component, but there limit much improvements gained using classical strategies. Generative adversarial networks (GAN) have been demonstrated generate impressive results, also successful as tool augmentation, mostly images limited diversity, medical...
Despite its benefits for tumour detection and treatment, the administration of contrast agents in dynamic contrast-enhanced MRI (DCE-MRI) is associated with a range issues, including their invasiveness, bioaccumulation, risk nephrogenic systemic fibrosis. This study explores feasibility producing synthetic enhancements by translating pre-contrast T1-weighted fat-saturated breast to corresponding first DCE-MRI sequence leveraging capabilities generative adversarial network (GAN)....