News

MedUni Vienna's Institute of Pathophysiology and Allergy Research has been named a Centre of Excellence for Quantitative Digital Microscopy. The research facility at MedUni Vienna's Centre for Pathophysiology, Infectiology and Immunology has thus been recognised by the manufacturer TissueGnostics as a global leader in the use of this technology, known as the TissueFAXS platform.

Link: https://www.meduniwien.ac.at/web/ueber-uns/news/2022/news-im-oktober-2022/institut-fuer-pathophysiologie-und-allergieforschung-zum-exzellenzzentrum-fuer-quantitative-digitale-mikroskopie-ernannt/

Medical University Vienna, represented by Isabella Ellinger and Amirreza Mahbod from Insitute for Pathophysiology and Allergy Research, Georg Dorffner from Institute for Artificial Intelligence and Adi Ellinger, together with a team from TissueGnostics GmbH (Rudolph Jedletzberger, Leon Kirsanov, Karly Briffa, Anastasiia Marchuk) presented their common FFG-funded project "Deep Learning-based knowledge transfer methods for nuclei segmentation in microscopic images".

Festival link: https://wirtschaftsagentur.at/veranstaltungen/wiener-forschungsfest-2022-1504/

Amirreza Mahbod presented an accepted paper (poster presentation) at the ICPR conference.

Conference Link: https://www.icpr2022.com/

Paper links: https://ieeexplore.ieee.org/abstract/document/9956253

Amirreza Mahbod presented his recent study at the MIC 2022 festival.

Title: Investigating the Impact of the Bit Depth of Fluorescence-Stained Images on the Performance of Deep Learning-Based Nuclei Instance Segmentation

Link: https://www.mdpi.com/2075-4418/11/6/967

Marcel Koseler presented his latest results from the master thesis at the17th YSA Symposium.

Title: Improving Generalisation Capability of Deep Learning-Based Nuclei Instance Segmentation Model

We participated with the station "Artificial Intelligence for the Analysis of Histological Images" (station 33 of the medical research mile in the area "Medical Imaging & Diagnostics")

An interdisciplinary team consisting of employees from Isabella Ellinger (IPA), Georg Dorffner (Institute for Artificial Intelligence) and the company TissueGnostics GmbH informed the numerous visitors to this station with digital presentations and many games on the topic "Analysis of histological images in clinic and research" and presented the results of the joint, FFG-funded Bridge Young Scientists research project "Deep Learning-based knowledge transfer methods for nuclei segmentation in microscopic images" (2020-2022).

Amirreza Mahbod is invited to give a presentation on "Deep Learning for Histological Image Analysis"

Registration link: https://us06web.zoom.us/webinar/register/WN_szlt92yER2uOX_N4K9XxDg

Amirreza Mahbod and Isabella Ellinger are serve as guest editors for a new special issue of Diagnostics journal (ISSN 2075-4418. impact factor: 3.706).

Link to the special issue: https://www.mdpi.com/journal/diagnostics/special_issues/Computer_Aided_Nuclei_Histological

On 15 March 2022, Benjamin Banker successfully defended his master thesis entitled: "Improving Mask R-CNN for nuclei Instance segmentation in H&E-stained images" and got the best mark (i.e. Sehr gut in Austrian grading system).

Isabella Ellinger was invited to the ÖQUASTA symposium and had an oral presentation about nuclei segmentation in microscopic images.

Symposium link: https://www.oequasta-symposium.com/events/symposium-2021/

Amirreza Mahbod et al. method for MICCAI 2021Foot Ulcer Segmentation challenge achieved an average Dice score of 88.8% and placed at rank 1 in the challenge leaderboard.

Leaderboard link: https://uwm-bigdata.github.io/wound-segmentation/

Challenge links: https://fusc.grand-challenge.org/

https://github.com/uwm-bigdata/wound-segmentation/tree/master/data/Foot%20Ulcer%20Segmentation%20Challenge

Amirreza Mahbod presented two accepted papers (one at the main conference and one at the AIHA workshop) at the ICPR conference. These works were related to microscopic image patch classification.

Congress Link: https://www.icpr2020.it/

Papers (preprints) links:

• https://arxiv.org/abs/2006.14715

• https://arxiv.org/abs/2011.07428

Amirreza Mahbod and Benjamin Bancher succesfully presented their recent works at the MIC 2020 festival.

Amirreza's work was selected as one of the top 6 abstracts for a rapid fire presentation at the congress.

Congress Link: https://www.meduniwien.ac.at/web/ueber-uns/events/2020/mic-festival-2020/

• Transfer Learning Using a Multi-Scale and Multi-Network Ensemble for Skin Lesion Classification, Mahbod A , Schaefer G , Wang C, Dorffner G , Ecker R , Ellinger I;

• Improved Mask R-CNN for Nuclei Segmentation in Histologic Images, Benjamin Bancher, Amirreza Mahbod, Isabella Ellinger, Georg Dorffner

Lists of academic groups and companies that are working on computer-based digital pathology image analysis (maintained by Amirreza Mahbod)

Link on GitHub: https://github.com/masih4/DigiPathoImageAnalysis

Our dataset of nuclei instance segmentation of digitized frozen H&E-stained histological image patches is now publicly available on Kaggle. Besides the dataset, standard U-net model codes (with and without post-processing) and also the exploratory data analysis (EDA) are also available on the kernel section.

Link on Kaggle: https://www.kaggle.com/ipateam/segmentation-of-nuclei-in-cryosectioned-he-images

Amirreza Mahbod et al. method for multi-organ nuclei segmentation and classification achieved an average PQ score of 56.07% and placed at rank 1 in the MoNuSAC post-challenge leaderboard.

Link for further details: https://monusac-2020.grand-challenge.org/Results/

Aim of the project is to generate and release the first annotated data set of digitized frozen H&E-stained histological sections from various organs enabling to train and validate algorithms for nuclei instance segmentation in different human organs.

Link for further details: https://www.kaggle.com/open-data-research-grant-2020-awardees?utm_medium=social&utm_source=twitter&utm_campaign=open-data-awardees-social