Projects & Software

Current research projects

BIGPICTURE

The goal of Bigpicture is to accelerate the development of AI in pathology by providing a large repository of high-quality annotated pathology data, accessible in a responsible, inclusive and sustainable way.

Deep Derma

The aim of this project is to apply artificial intelligence to detect basal cell carcinoma.

Deep PCA

In this project, we will combine deep learning and digitized whole-slide imaging of prostate cancer for reproducible extraction of quantitative biomarkers.

DIAGGRAFT

Investigation of AI for kidney transplant pathology.

IGNITE

The goal of IGNITE is to use automatic biomarker extraction with deep learning to predict the response of non-small cell lung cancer patients to immunotherapy.

PANCAIM

Investigation of AI for pancreatic cancer in radiology, pathology, genomics.

PathRad Fusion

The aim of this project is to integrate histopathological and radiological images to improve our understanding of disease diagnosis and progression in prostate cancer.

STIC

The aim of STIC is to improve the diagnostics of precursor lesions to high grade serous carcinoma (HGSC), the most common and lethal form of ovarian cancer.

Tumor budding

In this project, we will develop and validate digital image analysis algorithms for quantification of tumor budding from scanned whole slide images.

UNESCO

The aim of UNESCO is to study techniques for estimating uncertainty in computational pathology.

UNIC

The aim of UNIC is to develop artificial intelligence methods to refine diffuse-type gastric cancer (DGC) diagnostics.

ASAP - Fluid whole-slide image viewer

ASAP (Automated Slide Analysis Platform) is a fast and fluid viewer for digitized multi-resolution histopathology images. ASAP offers several tools to make annotations in an intuitive way. Dots, rectangles, polygons are all supported. ASAP allows on-slide visualization of image analysis and machine learning results such as segmentation masks with customizable lookup-tables.

More information Github repository

Automated Gleason Grading

Automated analysis of prostate biopsies following the Gleason grading system.

More information

HookNet

HookNet: multi-resolution convolutional neural networks for semantic segmentation in histopathology whole-slide images

More information Github repository

Neural Image Compression

Neural image compression for gigapixel histopathology image analysis.

More information Github repository

Finished projects

AMI

Automation in Medical Imaging - a joint project between DIAG and Fraunhofer MEVIS aimed at the development of a generic platform for automatic medical image analysis.

AQUILA

The goal of AQUILA is to investigate the prognostic value of Tumor Infiltrating Lymphocytes (TILs) in breast and colon cancer.

Artifact detection in digitized histopathology images

Development of a deep learning algorithm that can classify the different types of artifacts in whole slide images.

CAMELYON16

ISBI 2016 challenge on cancer metastases detection in lymph node.

CAMELYON17

ISBI 2017 challenge to evaluate algorithms for automated detection and classification of breast cancer metastases in whole-slide images of histological lymph node sections.

DCIS

The aim of this project is to use deep learning for histological assessment of the stroma for improved risk stratification of ductal carcinoma in situ (DCIS) patients.

ExaMode

The aim of ExaMode is to collect training data with limited human interaction for the processing of exascale volumes of healthcare data.

GENERATOR: Optimal treatment prediction for early stage lung cancer

We aim to combine advanced CT and FDG-PET image analysis with computational pathology to predict the most optimal treatment for each individual patient.

Identification of features in benign breast disease biopsies that predict breast cancer risk

Development of a deep learning system to predict BC risk in H&E

Multimot

MULTIMOT aims to build an open data ecosystem for cell migration research, through standardization, dissemination and meta-analysis efforts.

PANDA Challenge

Challenge on prostate cancer grading of biopsies using the Gleason grading system.

PRIM4BC

Detecting biomarkers for improved prognosis for triple negative breast cancer by combining histopathology, multiplex immunohistochemistry and Deep Learning.

PROACTING

The aim of PROACTING is to predict neoadjuvant chemotherapy treatment response from a single pre-operative core-needle biopsy of breast cancer tissue.

STITPRO I

Project on the development of ASAP and on the automatic detection of metastases in WSI of sentinel lymph nodes of breast cancer patients.

STITPRO II

Project with focus on the extension of open-source software ASAP that includes functionality for study and case management.

SysMIFTA

The investigation of the role of immune cell subsets in interstitial fibrosis and tubular atrophy in renal allografts, using multiplex immunohistochemistry and Deep Learning.

TIGER Challenge

Challenge on tumor infiltrating lymphcytes

Automated Quantification of Tumor-Infiltrating Lymphocytes

Developing an algorithmn that can automatically detect and segment tumor-infiltrating lymphocytes in breast cancer.

Tumor-Stroma Ratio

We aim at developing automatic and reproducible quantification of Tumor-Stroma Ratio in Whole-Slide Images using Deep Learning.

VPH-PRISM

Study on the correlations between several radiological modalities and histpathology for more understanding of breast cancer presentation.