First Person Vision for Activity Prediction Using Probabilistic Modeling

Abstract

Identifying activities of daily living is an important area of research with applications in smart-homes and healthcare for elderly people. It is challenging due to reasons like human self-occlusion, complex natural environment and the human behavior when performing a complicated task. From psychological studies, we know that human gaze is closely linked with the thought process and we tend to “look” at the objects before acting on them. Hence, we have used the object information present in gaze images as the context and formed the basis for activity prediction. Our system is based on HMM (Hidden Markov Models) and trained using ANN (Artificial Neural Network). We begin with extracting motion information from TPV (Third Person Vision) streams and object information from FPV (First Person Vision) cameras. The advantage of having FPV is that the object information forms the context of the scene. When context is included as input to the HMM for activity recognition, the precision increases. For testing, we used two standard datasets from TUM (Technische Universitaet Muenchen) and GTEA Gaze+ (Georgia Tech Egocentric Activities). In the first round, we trained our ANNs only with activity information and in the second round added the object information as well. We saw a significant increase in the precision (and accuracy) of predicted activities from 55.21% (respectively 85.25%) to 77.61% (respectively 93.5%). This confirmed our initial hypothesis that including the focus of attention of the actor in the form of object seen in FPV can help in predicting activities better.

Authors and Affiliations

Shaheena Noor, Vali Uddin

Keywords

Related Articles

Effect of Bridge Pier Shape on Scour Depth at Uniform Single Bridge Pier

Bridge pier scouring may lead to the bridge failure and the shape of bridge pier itself is one of the main factor to control scouring around bridge pier. The amount of sediment which is removed from the boundary of bridg...

Millimeter Waves Frequency Reconfigurable Antenna for 5G Networks

5G (Fifth Generation) is the next generation of data network, offering faster speeds and reliable connections on smart phones and other devices than ever before. These networks are still under development and expected to...

Power Flow and Transient Stability Enhancement using Thyristor Controlled Series Compensation

TL (Transmission Line) congestion is a key factor that affects the power system operational cost. In addition of renewable generation in National Grid of Pakistan, transmission line congestion are frequent. Consequently,...

The (2n+1)^2-Point Scheme Based on Bivariate Quartic Polynomial

We are going to implement least squares approach to fit the bivariate quartic polynomial to (2n+1)2- perceptions/data, where n>2. By taking different values of n, (2n+1)2-point approximating subdivision schemes are built...

Effect of Firing on Cracking and Warping of Clay Beams

Reinforced baked clay beams may be considered to be a substitute of reinforced cement concrete beams in order to build low cost houses. The baking of these clay beams can pose problems such as cracking and warping. This...

Download PDF file
  • EP ID EP394637
  • DOI 10.22581/muet1982.1804.09
  • Views 101
  • Downloads 0

How To Cite

Shaheena Noor, Vali Uddin (2018). First Person Vision for Activity Prediction Using Probabilistic Modeling. Mehran University Research Journal of Engineering and Technology, 37(4), 545-558. https://europub.co.uk/articles/-A-394637