Publications: | |
|---|---|
|
3D Pose Based Feedback For Physical ExercisesZiyi Zhao, Sena Kiciroglu, Hugues Vinzant, Yuan Cheng, Isinsu Katircioglu, Mathieu Salzmann, Pascal Fua ACCV 2022 Unsupervised physical training can cause serious injuries if performed incorrectly. We introduce a learning-based framework that identifies the mistakes made by a user and proposes corrective measures for easier and safer individual training. |
|
Long Term Motion Prediction Using KeyposesSena Kiciroglu, Wei Wang, Mathieu Salzmann, Pascal Fua 3DV 2022 - Oral We predict diverse and realistic long term (5 second) human motion by making use of "keyposes", the set of poses which we can use to accurately reconstruct the complete sequence. |
|
Motion Prediction Using Temporal Inception ModuleTim Lebailly, Sena Kiciroglu, Mathieu Salzmann, Pascal Fua, Wei Wang ACCV 2020 We predict human motion accurately in the short and long term future by making use of a Temporal Inception Module (TIM). Using TIM, we produce input embeddings using convolutional layers, by using different kernel sizes for different input lengths. |
|
ActiveMoCap: Optimized Viewpoint Selection for Active Human Motion CaptureSena Kiciroglu, Helge Rhodin, Sudipta N. Sinha, Mathieu Salzmann, Pascal Fua CVPR 2020 - Oral Given a short video sequence, we introduce an algorithm that predicts which viewpoints should be chosen by a moving camera to capture future frames so as to maximize 3D human pose estimation accuracy. |