Center Symmetric Local Descriptors for Image Classification

Center Symmetric Local Descriptors for Image Classification

Vaasudev Narayanan (Indian Institute of Technology (Indian School of Mines), Dhanbad, India) and Bhargav Parsi (University of California, Los Angeles, USA)
Copyright: © 2018 |Pages: 15
DOI: 10.4018/IJNCR.2018100104
OnDemand PDF Download:
No Current Special Offers


Local feature description forms an integral part of texture classification, image recognition, and face recognition. In this paper, the authors propose Center Symmetric Local Ternary Mapped Patterns (CS-LTMP) and eXtended Center Symmetric Local Ternary Mapped Patterns (XCS-LTMP) for local description of images. They combine the strengths of Center Symmetric Local Ternary Pattern (CS-LTP) which uses ternary codes and Center Symmetric Local Mapped Pattern (CS-LMP) which captures the nuances between images to make the CS-LTMP. Similarly, the auhtors combined CS-LTP and eXtended Center Symmetric Local Mapped Pattern (XCS-LMP) to form eXtended Center Symmetric Local Ternary Mapped Pattern (XCS-LTMP). They have conducted their experiments on the CIFAR10 dataset and show that their proposed methods perform significantly better than their direct competitors.
Article Preview

Literature Review

We present here a brief review of the various methods for texture classification that are proposed in the literature. Ojala, Pietikainen and Harwood (1996) introduced the Local Binary Pattern (LBP) as an approach for texture classification. Owing to its success in the texture classification task, the LBP operator then finds its way into various classification problems such as facial recognition as proposed in Ahonen, Hadid and Pietikainen (2006). Rotation invariance is something the LBP operator does not possess which is a disadvantage of this operator. Thus, in Ojala, Pietikainen, and Maenpaa (2002) a rotation invariant extension of LBP was introduced. Later it was found out that LBP was sensitive to global intensity variations and also to local intensity along edge components. To handle these shortcomings in Jun and Kim (2012) a new method was introduced called Local Gradient Pattern (LGP). LGP was shown to have higher discriminant power than LBP in the case of facial recognition.

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 11: 4 Issues (2022): Forthcoming, Available for Pre-Order
Volume 10: 4 Issues (2021): 3 Released, 1 Forthcoming
Volume 9: 4 Issues (2020)
Volume 8: 4 Issues (2019)
Volume 7: 4 Issues (2018)
Volume 6: 2 Issues (2017)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing