Volume 4 Number 3 (Jun. 2012)
Home > Archive > 2012 > Volume 4 Number 3 (Jun. 2012) >
IJCEE 2012 Vol.4(3): 260-263 ISSN: 1793-8163 DOI: 10.7763/IJCEE.2012.V4.491

A Novel Algorithm to Optimize Image Scaling in OpenCV: Emulation of Floating Point Arithmetic in Bilinear Interpolation

Samyuktha H. Subramanian
Abstract—Scaling is a complex algorithm wherein interpolation techniques are designed to achieve an optimum between factors such as efficiency and smoothness. One of these techniques is bilinear interpolation wherein the surrounding 2-by-2 source pixel values can be linearly weighed according to how close they are to the destination pixel. However, for applications in embedded systems with weak or no FPUs (floating point processors), the floating point values that arise after calculating weighted averages of the new pixel values have to be emulated. The algorithm presented to bypass floating point usage has been constructed based on the concept of resampling. It manipulates the pixels such that the fractional parts in the resulting floating numbers obtained are discrete in step, so that an optimal integer may be used to convert the floating point values to unsigned integers or signed integers by up-sampling. It has primarily been designed for scale factors, of integers which are multiples of 4 as well as the case of 2. This algorithm has been implemented in code in OpenCV and has been found to be faster and more efficient than the resize functions in OpenCV, and is therefore an alternative method to the scaling functions of the aforementioned library.

Index Terms—Image scaling, interpolation, resampling, opencv

S. H. Subramanian is with Birla Institute of Science and Technology, Pilani, India.

Cite: Samyuktha H. Subramanian, "A Novel Algorithm to Optimize Image Scaling in OpenCV: Emulation of Floating Point Arithmetic in Bilinear Interpolation," International Journal of Computer and Electrical Engineering vol. 4, no. 3, pp. 260-263, 2012.

General Information

ISSN: 1793-8163 (Print)
Abbreviated Title: Int. J. Comput. Electr. Eng.
Frequency: Quarterly
Editor-in-Chief: Prof. Yucong Duan
Abstracting/ Indexing: INSPEC, Ulrich's Periodicals Directory, Google Scholar, EBSCO, ProQuest, and Electronic Journals Library
E-mail: ijcee@iap.org

What's New

  • Jun 03, 2019 News!

    IJCEE Vol. 9, No. 2 - Vol. 10, No. 2 have been indexed by EI (Inspec) Inspec, created by the Institution of Engineering and Tech.!   [Click]

  • May 13, 2020 News!

    IJCEE Vol 12, No 2 is available online now   [Click]

  • Mar 04, 2020 News!

    IJCEE Vol 12, No 1 is available online now   [Click]

  • Dec 11, 2019 News!

    The dois of published papers in Vol 11, No 4 have been validated by Crossref

  • Oct 11, 2019 News!

    IJCEE Vol 11, No 4 is available online now   [Click]

  • Read more>>