DocumentCode :
702683
Title :
A novel algorithm for real time human classifier using Skin Colour identification
Author :
Bhat, Sandeep ; Meenakshi, Manjalagiri
Author_Institution :
Dept. of E&CE, Srinivas Inst. of Technol., Mangaluru, India
fYear :
2015
fDate :
8-10 Jan. 2015
Firstpage :
1
Lastpage :
5
Abstract :
This paper presents a novel approach for human classification based on skin colour identification technique. Next the same classifier is extended for the path planning and control of autonomous robots. The techniques for obstacle identification used in this work are Skin Colour Based (SCB), Pixel Count Based (PCB), Correlation Coefficient Based (CCB) and Histogram methods. In real- time obstacle detection, the Pixel Count Based (PCB) algorithm, Correlation Coefficient Based (CCB) and Skin Colour Based (SCB) algorithm are used. In this work CCB and PCB methods compare the similarities between two objects but SCB algorithm is to identify whether the tracked object is human or nonhuman in real time. Real time experimental results demonstrated the accuracy of CCB, PCB and SCB algorithms are 87.5% and 88.8% and 90.9% respectively and the time of CCB, PCB and SCB algorithms are 6.27sec, 6.50sec and 8.67sec respectively.
Keywords :
image classification; object detection; statistical analysis; CCB identification technique; PCB identification technique; SCB identification technique; autonomous robot control; correlation coefficient based identification technique; histogram method; object tracking; obstacle identification; path planning; pixel count based identification technique; realtime human classifier; skin colour based identification technique; skin colour identification technique; Algorithm design and analysis; Correlation coefficient; Feature extraction; Image color analysis; Object recognition; Real-time systems; Skin; Correlation Coefficient; Obstacle detection; Obstacle identification; Pixel Count; Skin Colour;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Pervasive Computing (ICPC), 2015 International Conference on
Conference_Location :
Pune
Type :
conf
DOI :
10.1109/PERVASIVE.2015.7087017
Filename :
7087017
Link To Document :
بازگشت