In “OpenCV Based Real-Time Video Processing Using Android,“ Ammar Anuar, Khairul Muzzammil Saipullah, and Nurul Atiqah Ismail from the Universiti Teknikal Malaysia at Melaka, with Yewguan Soo, Senior Lecturer at the University of Tokyo, describe two Android applications based on video processing, one using the OpenCV library and the other using Android library with self-implemented algorithm called CamTest.
Eight image processing methods are applied to each frame of the video captured from the Android smartphone. The smartphone used in this study is the Samsung Galaxy S, with Android 2.3 Gingerbread Operating System. The efficiencies and power consumptions of the two applications are compared by observing their frame processing rate and power consumption.
The experimental results show that out of the eight image processing methods, six methods that executed using OpenCV library are faster than that of CamTest with a total average ratio of 0.41. For the power consumption per frame test, six methods that executed using OpenCV library consume less power than that of CamTest with a total average ratio of 0.39.