Research Article
Human Object Classification Based on Nonsubsampled Contourlet Transform Combined with Zernike Moment
@INPROCEEDINGS{10.1007/978-3-319-29236-6_21, author={Luu Phuong and Nguyen Binh}, title={Human Object Classification Based on Nonsubsampled Contourlet Transform Combined with Zernike Moment}, proceedings={Context-Aware Systems and Applications. 4th International Conference, ICCASA 2015, Vung Tau, Vietnam, November 26-27, 2015, Revised Selected Papers}, proceedings_a={ICCASA}, year={2016}, month={4}, keywords={Object classification Zernike moment Nonsubsampled contourlet transform}, doi={10.1007/978-3-319-29236-6_21} }- Luu Phuong
Nguyen Binh
Year: 2016
Human Object Classification Based on Nonsubsampled Contourlet Transform Combined with Zernike Moment
ICCASA
Springer
DOI: 10.1007/978-3-319-29236-6_21
Abstract
The surveillance systems are more and more popular because of the security needs, but the traditional ones do not meet human’s expectation. This paper proposes the algorithm to classify objects mainly based on their contour property which are represented by the amplitude of zernike moment on nonsubsampled contourlet transform of a binary contour image. This feature shows promising results by just a simple association with the aspect ratio but gives high accuracy. The aspect ratio helps contour feature in case that the image is too blurred to extract the object’s contour. It also plays as a weak filter with nearly no more computational cost except for a division to support contour feature when applying gentle boost algorithm.