Selection of Relevant Parameters for Human Locomotion Unsupervised Classification
Journal Title: Memoirs of the Scientific Sections - Year 2018, Vol 0, Issue
Abstract
A method for the automatic selection of the most relevant parameters for human locomotion classification is proposed. A set of 36 statistical parameters extracted from video sequences showing three basic movement types is used. Because the unsupervised classification is based on the k-means clustering algorithm, the sets of relevant parameters are determined by applying binary optimization metaheuristics using a clustering evaluation measure as objective function. Considering that the objective function is multimodal, all combinations which maximize it are retained. The binary versions of Particle Swarm and Black Hole algorithms were modified to manage the multiple solutions of the optimization process. The experiments revealed that the Black Hole algorithm leads to better results, even if it is considered a simplified version of the Particle Swarm Algorithm.
Authors and Affiliations
Silviu-Ioan Bejinariu, Ramona Luca, Florin Rotaru
Keywords
Related Articles
- EP ID EP513157
- DOI -
- Views 74
- Downloads 0
Memoirs of the Scientific Sections – Series IV. A Brief Survey
This is a brief survey of the evolution of the publication Memoirs of the Scientific Sections, the first academic publication of Romania, issued since the year 1867 with the initial titulature Annals of the Academy. Ana...
The Late Mediaeval Necropolis of Lăpuşna (Hânceşti County, Republic of Moldova). Preliminary Anthropological Results
The osteological material was exhumed in 2013 by archaeologists from the Archaeology Department of the Cultural Heritage Institute of the Academy of Science of Moldova, in the mediaeval necropolis of Lăpuşna (Hânceşti Co...
A Brief History of the Library of the Romanian Academy, Branch of Iaşi
The paper makes a brief historical survey of the Library of the Romanian Academy – Branch of Iasi, laying stress on the difficult moments the librarians and scientists had to face in the dark years of the communist perio...
Pleasure, Imagination, Fear and Joy: Applied Themes in Nicole Oresme’s de Configurationibus
The present work explores the idea that concepts present in Nicole Oresme’s work, and rooted in the medieval revival of Corpus Aristotelicum, are reflected throughout active areas of research today. Nicole Oresme’s volum...
A Gis-Based Approach for Information Management in Geolinguistics
The geographic position is an important property of linguistic information in geolinguistics. This paper proposes a Geographic Information System (GIS) based framework for linguistic information management, consulting an...