3rd Workshop on Maritime Computer Vision (MaCVi) 2025: Challenge Results

📅 2025-01-17
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
Computer vision models for unmanned surface vessels (USVs) and underwater scenarios suffer from poor robustness, compounded by the absence of standardized evaluation protocols. Method: The MaCVi 2025 workshop organized a global competition to systematically benchmark over 700 technical submissions, establishing the first unified maritime computer vision evaluation framework. This framework integrates statistical analysis with qualitative assessment across core tasks—including object detection, image segmentation, and multimodal perception—and leverages publicly available datasets, standardized evaluation code, and a real-time leaderboard to ensure reproducibility. Contribution/Results: The initiative delivers the first open-source, authoritative maritime vision benchmark—comprising all data, reference implementations, and a live leaderboard—significantly enhancing model generalization and robustness in challenging aquatic environments (e.g., turbid water, dynamic lighting, occlusion). It lays the foundational infrastructure for standardization and advancement of maritime AI research.

Technology Category

Application Category

📝 Abstract
The 3rd Workshop on Maritime Computer Vision (MaCVi) 2025 addresses maritime computer vision for Unmanned Surface Vehicles (USV) and underwater. This report offers a comprehensive overview of the findings from the challenges. We provide both statistical and qualitative analyses, evaluating trends from over 700 submissions. All datasets, evaluation code, and the leaderboard are available to the public at https://macvi.org/workshop/macvi25.
Problem

Research questions and friction points this paper is trying to address.

Maritime Computer Vision
Unmanned Vessels
Underwater Technology
Innovation

Methods, ideas, or system contributions that make the work stand out.

Marine Computer Vision
Unmanned Marine Vessels
Underwater Environments
🔎 Similar Papers
No similar papers found.
Benjamin Kiefer
Benjamin Kiefer
Universität Tübingen
Deep LearningObject DetectionComputer Vision
L
Lojze vZust
University of Ljubljana
J
Jon Muhovivc
University of Ljubljana
Matej Kristan
Matej Kristan
Full Professor at Faculty of computer and information science, University of Ljubljana
Computer visionMachine learningPattern recognition
J
Janez Pervs
University of Ljubljana
M
Matija Tervsek
Luxonis
U
Uma Mudenagudi Chaitra Desai
Center of Excellence in Visual Intelligence, KLE Technological University
Arnold Wiliem
Arnold Wiliem
Principal Engineer in Deep Learning, Shield AI - Vision Systems group; Adjunct A/Professor QUT
Deep LearningComputer VisionPattern RecognitionImage Processing
M
Marten Kreis
University of Tuebingen
N
Nikhil Akalwadi
Center of Excellence in Visual Intelligence, KLE Technological University
Y
Yitong Quan
University of Tuebingen
Zhiqiang Zhong
Zhiqiang Zhong
MSCA Postdoc & YIA Fellow at University of Luxembourg
Graph Machine LearningFoundation ModelAI4Science
Z
Zhe Zhang
Nanjing University of Science and Technology
S
Sujie Liu
Nanjing University of Science and Technology
X
Xuran Chen
Nanjing University of Science and Technology
Y
Yang Yang
Nanjing University of Science and Technology
M
Matej Fabijanić
University of Zagreb Faculty of Electrical Engineering and Computing
Fausto Ferreira
Fausto Ferreira
Assistant Professor, University of Zagreb Faculty of Electrical Engineering and Computing, Croatia
marine roboticsunderwater roboticsrobotics competitionsmarine law
S
Seongju Lee
Gwangju Institute of Science and Technology (GIST)
J
Junseok Lee
Gwangju Institute of Science and Technology (GIST)
K
Kyoobin Lee
Gwangju Institute of Science and Technology (GIST)
Shanliang Yao
Shanliang Yao
Yancheng Institute of Technology
Autonomous DrivingIntelligent VehiclesRadar-Camera FusionMaritime Perception
Runwei Guan
Runwei Guan
Hong Kong University of Science and Technology (Guangzhou) / Founder of FertiTech AI
Multi-Modal LearningUnmanned Surface VesselRadar PerceptionAI Medicine
X
Xiaoyu Huang
University of Liverpool
Y
Yiqian Ni
Xi’an Jiaotong-Liverpool University
H
Himanshu Kumar
Independent Researcher
Y
Yuan Feng
Dalian Maritime University, School of Marine Engineering
Yi-Ching Cheng
Yi-Ching Cheng
National Cheng Kung University
T
Tzu-Yu Lin
National Cheng Kung University
Chia-Ming Lee
Chia-Ming Lee
National Yang Ming Chiao Tung University
Computer VisionImage ProcessingInformation ForensicsMultimedia
Chih-Chung Hsu
Chih-Chung Hsu
Associate Professor of Institute of Intelligent Systems, College of AI, NYCU
Deep learningImage processingcomputer visionimage compressionvideo editing
J
Jannik Sheikh
Fraunhofer IOSB
A
A. Michel
Fraunhofer IOSB
W
Wolfgang Gross
Fraunhofer IOSB
Martin Weinmann
Martin Weinmann
Karlsruhe Institute of Technology (KIT)
Computer VisionPattern RecognitionMachine LearningPhotogrammetryRemote Sensing
J
Josip vSarić
University of Ljubljana
Y
Yipeng Lin
Nanjing University of Science and Technology
X
Xiang Yang
Nanjing University of Science and Technology
N
Nan Jiang
Nanjing University
Y
Yutang Lu
Beijing University of Posts and Telecommunications
Fei Feng
Fei Feng
Beijing University of Posts and Telecommunications
Ali Awad
Ali Awad
Grad. Student, Applied Computing, Michigan Technological University
RoboticsMachine-LearningControlEmbedded Systems
Evan Lucas
Evan Lucas
Michigan Technological University
A
Ashraf Saleem
Michigan Technological University
Ching-Heng Cheng
Ching-Heng Cheng
National Cheng Kung University
Y
Yufeng Lin
National Cheng Kung University