On this page you can find some selected publications with relative resources. For a complete list of papers, you can check this link or my Google Scholar profile.
Journal papers
Journal paper
Visual tracking in camera-switching outdoor sport videos: Benchmark and baselines for skiing
Matteo Dunnhofer and Christian Micheloni
Computer Vision and Image Understanding (2024)
Extended version of our WACV 2024 paper about tracking skiers in camera-switching broadcasting videos.
Deep convolutional feature details for better knee disorder diagnoses in magnetic resonance images
Matteo Dunnhofer, Niki Martinel, and Christian Micheloni
Computerized Medical Imaging and Graphics (2022)
This paper extends our MIDL 2021 work about employing knee-specific CNN architectures for a better extraction of features related to knee anomalies. Here we present also a model interpretation strategy.
Weakly-Supervised Domain Adaptation of Deep Regression Trackers via Reinforced Knowledge Distillation
Matteo Dunnhofer, Niki Martinel, and Christian Micheloni
IEEE Robotics and Automation Letters (2021)
We propose the first solution to adapt the generic knowledge of deep regression trackers to particular and small data vision domains. Reinforcement learning is used to express weak supervision and knowledge distillation to maintain learning stability.
Siam-U-Net: encoder-decoder siamese network for knee cartilage tracking in ultrasound images
Matteo Dunnhofer, Maria Antico, Fumio Sasazawa, Yu Takeda, Saskia Camps, Niki Martinel, Christian Micheloni, Gustavo Carneiro, and Davide Fontanarosa
Medical Image Analysis (2020)
This study proposes a new deep learning method to track, accurately and efficiently, the femoral condyle cartilage in ultrasound sequences, which were acquired under several clinical conditions, mimicking realistic surgical setups.
Matteo Dunnhofer, Luca Sordi, Niki Martinel, and Christian Micheloni
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2024
In this paper, we perform a study over visual object tracking algorithms to localize skiers in multi-camera videos capturing their complete performance, from the top to the bottom of the course.
CoCoLoT: Combining Complementary Trackers in Long-Term Visual Tracking
Matteo Dunnhofer and Christian Micheloni
International Conference on Pattern Recognition (ICPR) 2022
In this paper, we present a long-term tracking methodology that combines the complementary capabilities of different trackers to achieve more robust tracking performance. We used this solution for winning the VOT2021 long-term challenge.
The Ninth Visual Object Tracking VOT2021 Challenge Results
Matej Kristan, Jiří Matas, Aleš Leonardis, Michael Felsberg, Roman Pflugfelder, Joni-Kristian Kämäräinen, Hyung Jin Chang, Martin Danelljan, ..., Matteo Dunnhofer, ...
IEEE/CVF International Conference on Computer Vision (ICCV) 2021, Visual Object Tracking Challenge VOT2021 workshop
This paper surveys the state-of-the-art in visual object tracking for year 2021. It includes the description of our winning solution to the long-term challenge mlpLT.
Improving MRI-based Knee Disorder Diagnosis with Pyramidal Feature Details
Matteo Dunnhofer, Niki Martinel, and Christian Micheloni
International Conference on Medical Imaging with Deep Learning (MIDL) 2021
We present a plug-and-play CNN architecture to better extract features related to knee anomalies when imaged with MRI. Our strategy allows to improve the diagnostic performance of baseline CNNs.
Tracking-by-Trackers with a Distilled and Reinforced Model
Matteo Dunnhofer, Niki Martinel, and Christian Micheloni
Asian Conference on Computer Vision (ACCV) 2020
This paper proposes a novel tracking methodology that takes advantage of other visual trackers, offline and online, and extensive validation shows that the proposed algorithms compete with real-time state-of-the-art trackers.