Roland Kwitt

Full Professor
Department of Artificial Intelligence and Human Interfaces
University of Salzburg

View My GitHub Profile

Publications - Teaching - Contact - Blog


I am a full professor for machine learning in the Department of Artificial Intelligence and Human Interfaces (AIHI) at the University of Salzburg (PLUS), Austria (and currently deputy head of the department). Prior to that, I was part of the medical imaging and computer vision group at Kitware Inc., North Carolina, USA. My research spans multiple areas, but mostly focusses on theoretical and practical aspects of learning methods that allow to leverage and control structural characteristics of data. I am also a member of the ELLIS society.

News

Research interests

Current group members

Past group members

Selected publications

For a complete list of all papers, go to Publications, or visit my Google Scholar profile.

Connectivity-Optimized Representation Learning via Persistent Homology
ICML 2019
Feature-Space Transfer for Data Augmentation
CVPR 2018 (Oral)
Deep Learning With Topological Signatures
NIPS 2017
AGA: Attribute-Guided Augmentation
CVPR 2017 (Oral)
Regression Uncertainty on the Grassmannian
AISTATS 2017 (Oral)
One-Shot Learning of Scene Locations via Feature Trajectory Transfer
CVPR 2016 (Spotlight)
Parametric Regression on the Grassmannian
IEEE Trans. Pattern Anal. Mach. Intell. 2016
Statistical Topological Data Analysis - A Kernel Perspective
NIPS 2015
A Stable Multi-Scale Kernel for Topological Machine Learning
CVPR 2015
Scene Recognition on the Semantic Manifold
ECCV 2012
Recognition in Ultrasound Videos: Where Am I?
MICCAI 2012 (Oral)
Lightweight Probabilistic Texture Retrieval
IEEE Trans. Image Process. 2010

Short biography

Curriculum Vitae (CV)

Professional service

I regularly serve as a reviewer for the major vision and machine learning conferences (e.g., CVPR, ICCV, ECCV, BMVC, NIPS, AISTATS, ICIP, ICPR, MICCAI, ICLR) as well as the major journals in the field (e.g.,TPAMI, JMLR, IJCV, TMI, TIP, TSP, MedIA).

Awards