Human Activity Recognition Using Gaussian Mixtures

Abstract

Objective of recognizing human activity in the growing ubiquity and mobility of video data has emerged as one of the key research topic in computer vision and machine learning. This paper considers the inclusivity of latent variables by extracting missing information with discriminative features in reduced feature space. Multi-modal distribution analysis is performed using Gaussian mixture model in maximum likelihood basis to obtain Gaussian mixtures, PCA is applied to these mixtures to preserve suitable discriminatory features in compressed space. Different distance measure techniques are used to classify and recognize human activity. Proposed subspace mixture model achieved most promising results on KTH dataset in comparison with few state-of-the-art techniques.

Authors and Affiliations

Mahantesh K, Namita H, Harshitha R, Madhurya S, Nischitha K S

Keywords

Related Articles

3 D Printing: Techniques and Applications

We will firstly elaborate the meaning of 3D printing and its technologies to look out the processes and material in different industries like manufacturing or production. 3D printing has become crucial and embryonic sec...

Secure Data Deduplication in Cloud

Secure deduplication is a system for discarding duplicate copies of limit data, and offers security to them. To lessening storage space and exchange information move limit in circulated stockpiling deduplication has bee...

Solar Panel With Wind Turbine Power Generation System

In this project we generate an electric power through the fabrication of horizontal axis wind mill and pyramid shaped solar panel. A wind turbine system is a machine that converts the kinetic energy in wind into mechani...

Fetal ECG Extraction & Analysis Using Wavelet Transform

The fetal electrocardiogram (ECG) is an analytical tool that measures and records the electrical movement of the heart of the fetus during pregnancy and provides fine details. Analysis of these details allows analysis o...

Improvement of Thermal Hydraulic Characteristics of MPFHS with Nanofluids –A Review

Today’s several new techniques have to be compelled to be explored for the developments of recent electronic devices for dissipating high heat flux for ensure stable and optimum operation. Like micro-channel heat sinks...

Download PDF file
  • EP ID EP24708
  • DOI -
  • Views 399
  • Downloads 13

How To Cite

Mahantesh K, Namita H, Harshitha R, Madhurya S, Nischitha K S (2017). Human Activity Recognition Using Gaussian Mixtures. International Journal for Research in Applied Science and Engineering Technology (IJRASET), 5(6), -. https://europub.co.uk/articles/-A-24708