Trajectory-based Descriptors for Action Recognition in Real-world Videos

Trajectory-based Descriptors for Action Recognition in Real-world Videos

This thesis explores motion trajectory-based approaches to recognize human actions in real-world, unconstrained videos. Recognizing actions is an important task in applications such as video retrieval, surveillance, human-robot interactions, analysis of sports videos, summarization of videos, behavi...

Full description

Bibliographic Details
Main Author: Narayan, Sanath
Other Authors: Ramakrishnan, K R
Language:en_US
Published: 2018
Subjects:
Trajectory-based Descriptors
Action Recognition
Egocentric Videos
Covariance Descriptors
Hyper-Fisher Vector
Dense Trajectories
Fisher Vectors
Fisher Kernel
Fisher Vector (FV) Coding Method
Motion Segmentation Method
Electrical Engineering
Online Access:http://etd.iisc.ernet.in/2005/3914
http://etd.iisc.ernet.in/abstracts/4791/G27146-Abs.pdf

Internet

http://etd.iisc.ernet.in/2005/3914
http://etd.iisc.ernet.in/abstracts/4791/G27146-Abs.pdf

Similar Items