DATA COMPETITIONS
BODYinTRANSIT
The BODYinTRANSIT Data Competition is part of the ERC-funded BODYinTRANSIT project and is hosted at the IEEE Machine Learning for Signal Processing (MLSP) Workshop 2025. This competition is dedicated to the decoding of human body perception through multimodal biosignals, thereby challenging participants to develop classification models that accurately predict perceived changes in body weight (lighter, heavier, or unchanged) based on physiological and kinematic time-series data. By participating in this competition, researchers will contribute significantly to the advancement of signal processing methodologies pertinent to biosignal analysis and will also acquire invaluable insights concerning human sensorimotor integration.
Organizers:
Mohammad Mahdi Dehshibi, Universidad Carlos III de Madrid, Spain
Tomás Martínez Cortés, Universidad Carlos III de Madrid, Spain
Fernando Díaz-de-María, Universidad Carlos III de Madrid, Spain
Nadia Bianchi-Berthouze, University College London, United Kingdom
Ana Tajadura-Jiménez, Universidad Carlos III de Madrid, Spain
For more information please click here
Non-native Children's Automatic Speech Assessment Challenge (NOCASA)
Learning the pronunciation of foreign or second language (L2) requires a lot of practise and accurate feedback. Mobile apps that have automatic pronunciation assessment (APA) technology lets learners practise pronunciation at their own time, place and pace. Challenges: The developing and implementing APA has several challenges. First and worst of all is the lack of speech data for L2 learners that would be annotated for pronunciation accuracy. This is particularly the case of children and the learners of low-resource target languages. Second, if such data were available, it is usually heavily unbalanced for different skill levels and the provided reference scores suffer from noise and inter-annotator disagreement. Finally, to make a useful app, the scoring has to happen in seconds to provide real-time feedback with minimal delay to encourage a lot of repetitions.
Data: Recordings of second language learning children of 5 - 12 years repeating the words in Norwegian that were played to them. For each word in the data we provide the correct orthographic transcription and the speech accuracy assessment score of 1 - 5 given by human experts.
Task: Develop an automatic speech assessment system and use it to predict the score of each utterance in the given test data.
Organizers:
Mikko Kurimo (mikko.kurimo[at]aalto.fi), Aalto University, Finland,
Giampiero Salvi (giampiero.salvi[at]ntnu.no), NTNU, Norway,
Tamás Grósz (tamas.grosz[at]aalto.fi) Aalto University, Finland,
Sari Ylinen (sari.p.ylinen[at]tuni.fi) Tampere University, Finland,
Minna Lehtonen minna.h.lehtonen[at]utu.fi University of Turku, Finland,
Sofia Strömbergsson (sofia.strombergsson[at]ki.se) Karolinska Institutet, Sweden,
Torbjørn Svendsen (torbjorn.svendsen[at]ntnu.no) NTNU, Norway
For more information please click here
Sampling-Assisted Pathloss Radio Map Prediction
The Sampling-Assisted Pathloss Radio Map Prediction Data Competition at MLSP 2025 invites participants to develop deep learning models for predicting indoor pathloss radio maps using indoor geometry (floor plans), electromagnetic properties of building materials (reflectance and transmittance), and transmitter locations. The challenge includes two tasks that integrate these inputs with sparse pathloss samples: (1) predicting radio maps augmented by uniformly sampled ground-truth measurements; and (2) exploring sampling strategies for radio map estimation, where participants design approaches for selecting measurement locations. This challenge focuses on advancing sample-aided data-driven models for radio map prediction using a ray-tracing generated dataset that simulates various indoor propagation scenarios, building on previous challenges while emphasizing the role of sparse spatial samples in improving prediction accuracy
Organizers:
Çagkan Yapar,
Stefanos Bakirtzis,
Andra Lutu,
Ian Wassell,
Jie Zhang,
Giuseppe Caire
For more information please click here
VEELA - Vessel Extraction and Extrication for Liver Analysis
The VEELA challenge aims to initiate the development of advanced hepatic and portal vessel segmentation techniques from computed tomography angiography (CTA) images of healthy people (i.e., living donated liver transplantation candidates) and the design of new algorithms that can work under varying image quality and contrast levels.
The VEELA challenge is being organized as a successor to the CHAOS challenge (chaos.grand-challenge.org/) VEELA deepens the subject as it moves from solid organ segmentation to inner vascular tree extraction. To increase the connection between CHAOS and VEELA, the VEELA training and test datasets are chosen to be identical to the CHAOS training and test datasets.
Organizers:
M. Alper Selver
Oğuz Dicle
N. Sinem Gezer
İlker Özgür Koska
Hazım Kemal Ekenel
Ilkay Oksuz
Ziya Ata Yazıcı
Pierre-Henri Conze
Tugce Toprak Tepegoz
Emre Kavur
Pervin Bulucu