Gesture Modeling/Action Recognition
In collaboration with researchers at BU and Northeastern University, we have developed a method for gesture modeling based on Hanklet.
Our method decomposes a gesture into sub-trajectories that are the output of a sequence of atomic linear time invariant (LTI) systems. We use a Hidden Markov Model to model the transitions from the LTI system to another.
We describe the set of trajectories in a temporal window by the corresponding Hankel matrix (Hanklet) and train a set of HMMs (one for each gesture class) with a discriminative approach.
The code of our implementation will be made publicly available very soon. If you need of any information, please email me!
You may be interested in the following paper:
- Lo Presti L., La Cascia M., Sclaroff S., Camps O., “Gesture Modeling by Hanklet-based Hidden Markov Model“, Asian Conference on Computer Vision (ACCV) 2014, Singapore (pdf)
- Lo Presti L., La Cascia M., Sclaroff S., Camps O., “Hankelet-based Dynamical Systems Modeling for 3D Action Recognition,” Elsevier Image and Vision Computing (IMAVIS), 2015 (pdf)
Computational Behavioral Science
During my post-doc at Boston University (BU), I have been involved in a great project whose goal is the development of methods for analyzing the social behavior of children interacting with adults. Among the other things, the project aims at finding methods to measure if and how children improve their social and communicative skills.
In collaboration with researchers at BU and GeorgiaTech, I have studied methods to capture and represent the behavioral signals from videos acquired by a camera network. I have implemented methods to measure the “engagement” of the children during a dyadic interaction and studied how behavioral reciprocity may help engagement prediction.
You may be interested in the following papers:
- Rehg J., Abowd G., Rozga A., Romero M., Clements M., Sclaroff S., Essa I., Ousley O., Li Y., Kim C., Rao H., Kim J., Lo Presti L., Zhang J., Lantsman D., Bidwell J., Ye Z., “Decoding Children’s Social Behavior,” IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), Portland, Oregon USA, June 2013. (pdf)
- Lo Presti L., Sclaroff S., Rozga A., “Joint Alignment and Modeling of Correlated Behavior Streams“, accepted to IEEE ICCV-Workshop on Decoding Social Cues from Social Interactions, Sydney, Australia Dec. 2013. (pdf)