MANO | Activities

We are excited to share that our Anomaly Detection Workshop (VAND 4.0) will take place at CVPR 2026! 
Already looking forward to your paper submissions and challenge participation.

https://sites.google.com/view/vand4-cvpr2026 

LesiOnTime, developed within the EUCAIM project, is a 3D deep-learning model designed to mirror radiologists’ real workflow on breast MRI by integrating temporal context and clinical reasoning.

It introduces:

• Temporal Prior Attention Block — compares previous and current MRI scans to capture subtle interval changes
• BI-RADS Consistency Loss — embeds clinical scoring directly into model training for more robust segmentation

By combining temporal information with clinically aligned intelligence, LesiOnTime advances AI-driven lesion detection in breast MRI.

• Paper: https://arxiv.org/pdf/2508.00496
• Code: https://github.com/cirmuw/LesiOnTime

We are happy to announce that the 3rd edition of the Visual Anomaly and Novelty Detection (VAND) Workshop will take place @CVPR2025!
We look forward to seeing you there :)
 

The event – covering several aspects concerning secondary use of data regarding the European Health Data Space (EHDS) Regulation, the Data Governance Act (DGA) and its implementation in Europe and Austria – included a panel discussion and three workshops.

We are happy to be part of the first Data Science Vienna Poster Workshop on Oct 1st 2024 at the Vienna Sternwarte. With over 60 registered participants it will be an inspiring event for the Viennese community.

Ein innovatives Projekt der Lehre an der MedUniWien wurde beim diesjährigen ArsDocendi-Staatspreis mit dem Anerkennungspreis prämiert.

Das Projekt „Forschungszentrierte Kompetenzerwerbung durch Journal Clubs“ von Philipp Seeböck wurde in der Kategorie „Forschungsbezogene bzw. kunstgeleitete Lehre“ ausgezeichnet.

Today at CVPR 2024: Visual Anomaly Detection Workshop! 

Anomaly detection is crucial for identifying unusual patterns and potential issues in data, making it essential for applications in security, healthcare, and beyond. At the VAND 2.0 Workshop, co-organized by Philipp, top researchers and industry experts are gathering to share insights and innovations in visual anomaly and novelty detection.

Live from CVPR 2024, Seattle, WA
Dive into cutting-edge research, network with fellow professionals, and see how our lab is contributing to transformative advancements in anomaly detection.

#AnomalyDetection #AI #MachineLearning #CVPR2024 #Research #Innovation #CIRlab

April 2024: Our novel method about improving segmentation using anomaly detection got accepted at Medical Image Analysis Journal! 

"Anomaly guided segmentation: Introducing semantic context for lesion segmentation in retinal OCT using weak context supervision from anomaly detection"

June 7th: new paper on prediction of anomalies in high-risk breast cancer women with deep learning by Bianca Burger and colleagues published in European Radiology Experimental.