Embedded Files
FILOVI (Analizar el comportamiento de consumidores con medidas simultáneas de FIsiología, LOcalización y VIsión por computador). This project is a collaboration between the RoPert research lab from the University of Zaragoza, and BitBrain.
We use data extracted from different sensors to study customer/user behavior in different scenarios.
Recognition from wearable devices.
Recognition from wearable devices.
One of the main sources of information are wearable devices. The first result in this project is a method to analyze the video extracted from a chest-mounted camera to detect and locate the daily objects that the user is seeing and interacting with. We are working on novel strategies to make object detection more robust and efficient in videos recorded in challenging conditions.
Related results and publications:
Related results and publications:
Robust and efficient post-processing for video object detection.
A. Sabater, L. Montesano and A. C. Murillo. IEEE Int. Conf. on Intelligent Robots and Systems (IROS), 2020. [Paper][Code][Supplementary video]
Performance of object recognition in wearable videos.
A. Sabater, L. Montesano and A. C. Murillo, IEEE Int. Conf. on Emerging Technologies and Factory Automation (ETFA), 2019, pp. 1813-1820. [Paper (IEEExplore)].
Corresponding poster at Young Professionals and Students Forum. ETFA 2019. [Poster]
[Code][Sample video][Models]
Sample of bounding box predictions from an unlabeled video
Analyzing the objects in the scene can directly give good insight on the type of scene (e..g., room) where the user is.
Activity Recognition
Activity Recognition
Besides understanding the scene and objects on the video, it is essential to analyze the activities performed by the users. We are working on different challenges
Novel strategies to perform action recognition when training data is limited and noisy.
Related results and publications:
Related results and publications:
Domain and View-point Agnostic Hand Action Recognition
A Sabater, I Alonso, L Montesano, AC Murillo. IEEE/RSJ Int. Conf. on Intelligent Robots and Systems (IROS), IEEE Robotics and Automation Letters, 2021. [Pdf (arxiv)] [Code][Supplementary video]
One-shot action recognition in challenging therapy scenarios.
A. Sabater, L. Santos, J. Santos-Victor, A. Bernardino, L. Montesano and A. C. Murillo, Workshop on Learning from Limited and Imperfect Data (L2ID), Computer Vision and Pattern Recognition Workshops (CVPRW), 2021 [Paper (arxiv)][Code][Supplementary video]
Action recognition from Event cameras
Related results and publications:
Event transformer. a sparse-aware solution for efficient event data processing.
Sabater, Alberto, Luis Montesano, and Ana C. Murillo. IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) 2022. [Pdf] [Code] [Supplementary video]
Event Transformer+. A multi-purpose solution for efficient event data processing.
Sabater, Alberto, Luis Montesano, and Ana C. Murillo. IEEE Transactions on Pattern Analysis and Machine Intelligence (2023). [Pdf (arxiv)] [Code]
During this project, we are also working analyzing user activities from external cameras, to recognize and track users, or identify actions or commands performed by the user.
Related results and publications:
Related results and publications:
Fine-Grained Pointing Recognition for Natural Drone Guidance.
O. Barbed, P. Azagra, L. Teixeira, M. Chli, J. Civera, A. C. Murillo. Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition Workshops. 2020. [pdf, data&code]
Distributed Multi-Target Tracking in Camera Networks.
S. Casao, A. Naya, A. C Murillo, E.Montijano. IEEE Int. Conf. on Robotics and Automation (ICRA) 2021. [pdf] [video]
Page updated
Report abuse