Benjamin Filtjens Human Movement Intelligence

Research project

FOG-IT

A clinician-in-the-loop dashboard supporting AI-assisted freezing-of-gait annotation during standardized laboratory tasks.

Research Collaboration Movement in clinical practice
Role
Postdoc
Award / reference year
2021
Partners
KU Leuven
Funders
KU Leuven Industrial Research Fund
Wearable movement recordings are processed by a neural network, reviewed by an expert, and turned into verified freezing-of-gait annotations.

AI-assisted freezing-of-gait assessment: initial model annotations, expert review, and verified results.

Adapted from Yang et al. (2026), via VSC · CC BY 4.0 · View full-size image ↗

A clinician-in-the-loop dashboard for FOG assessment

Assessing freezing of gait (FOG) requires experts to review movement recordings and identify individual episodes. This is time-consuming and can lead to differences between raters, even during standardized laboratory tasks.

FOG-IT develops a clinician-in-the-loop dashboard to support this work. AI proposes annotations; clinicians review and correct them, retaining the final say. The research evaluates methods based on motion capture and wearable inertial sensors, including how models transfer between cohorts and can be adapted using reviewed annotations.

The AI-assisted annotation platform demonstrates this workflow as a proof of concept. AID-FOG extends this research towards monitoring freezing of gait at home.

Patent application

System for assessment of freezing of gait — pending patent application PCT/EP2026/065548 (2026). Role: co-inventor.

Related work

Publications

2024

Journal Article

Freezing of gait assessment with inertial measurement units and deep learning: effect of tasks, medication states, and stops

Yang PK, Filtjens B, Ginis P, Goris M, Nieuwboer A, Gilat M, Slaets P, Vanrumste B

Journal of NeuroEngineering and Rehabilitation 21 (1), 24, 2024

Citation

Yang PK, Filtjens B, Ginis P, Goris M, Nieuwboer A, Gilat M, Slaets P, Vanrumste B (2024). Freezing of gait assessment with inertial measurement units and deep learning: effect of tasks, medication states, and stops. Journal of NeuroEngineering and Rehabilitation 21 (1), 24, 2024.

2024

Journal Article

Automatic detection and assessment of freezing of gait manifestations

Yang PK, Filtjens B, Ginis P, Goris M, Nieuwboer A, Gilat M, Slaets P, Vanrumste B

IEEE Transactions on Neural Systems and Rehabilitation Engineering 32, 2699-2708, 2024

Citation

Yang PK, Filtjens B, Ginis P, Goris M, Nieuwboer A, Gilat M, Slaets P, Vanrumste B (2024). Automatic detection and assessment of freezing of gait manifestations. IEEE Transactions on Neural Systems and Rehabilitation Engineering 32, 2699-2708, 2024. https://doi.org/10.1109/TNSRE.2024.3431208

2024

Preprint

Multimodal freezing of gait detection: Analyzing the benefits of physiological data

Yang PK, Filtjens B, Ginis P, Goris M, Nieuwboer A, Gilat M, Slaets P, Vanrumste B

medRxiv. Published online October 27, 2024:2024.10.25.24315880

Preprint (2024); journal version published in 2025.

Citation

Yang PK, Filtjens B, Ginis P, Goris M, Nieuwboer A, Gilat M, Slaets P, Vanrumste B (2024). Multimodal freezing of gait detection: Analyzing the benefits of physiological data. medRxiv. Published online October 27, 2024:2024.10.25.24315880. https://doi.org/10.1101/2024.10.25.24315880

2023

Conference Paper

Automated freezing of gait assessment with deep learning and data augmentation from simulated inertial measurement unit data

Filtjens B, Yang PK, Goris M, Gilat M, Kempynck N, Ginis P, Nieuwboer A, Slaets P, Vanrumste B

2023 IEEE 19th international conference on body sensor networks (BSN), 1-4, 2023

Citation

Filtjens B, Yang PK, Goris M, Gilat M, Kempynck N, Ginis P, Nieuwboer A, Slaets P, Vanrumste B (2023). Automated freezing of gait assessment with deep learning and data augmentation from simulated inertial measurement unit data. 2023 IEEE 19th international conference on body sensor networks (BSN), 1-4, 2023.

2022

Journal Article

Automated freezing of gait assessment with marker-based motion capture and multi-stage spatial-temporal graph convolutional neural networks

Filtjens B, Ginis P, Nieuwboer A, Slaets P, Vanrumste B

Journal of NeuroEngineering and Rehabilitation 19 (1), 48, 2022

Citation

Filtjens B, Ginis P, Nieuwboer A, Slaets P, Vanrumste B (2022). Automated freezing of gait assessment with marker-based motion capture and multi-stage spatial-temporal graph convolutional neural networks. Journal of NeuroEngineering and Rehabilitation 19 (1), 48, 2022. https://doi.org/10.1186/s12984-022-01025-3

2021

Journal Article

Modelling and identification of characteristic kinematic features preceding freezing of gait with convolutional neural networks and layer-wise relevance propagation

Filtjens B, Ginis P, Nieuwboer A, Afzal MR, Spildooren J, Vanrumste B, Slaets P

BMC medical informatics and decision making 21 (1), 341, 2021

Citation

Filtjens B, Ginis P, Nieuwboer A, Afzal MR, Spildooren J, Vanrumste B, Slaets P (2021). Modelling and identification of characteristic kinematic features preceding freezing of gait with convolutional neural networks and layer-wise relevance propagation. BMC medical informatics and decision making 21 (1), 341, 2021.