Circular Hough Transform and Integral Intensity Projection for Computing Automatic Footprint Arch Index

Abstract

This paper demonstrates approaches for automatic computing footprint arch index by Circular Hough Transform and Intensity Projection methods. We applied k-mean clustering approach in order to classify the cluster of the color in the footprint image. Next, the image is converted to a binary image and divided into left and right side. The Circular Hough Transform method and flood-fill operation are then utilized to detect and to remove the toe prints, respectively. After that the positions of the anterior and the posterior aspect of the toeless foot prints are defined by using the horizontal integral intensity projection method. Subsequently, an angle of the toeless foot print is then estimated from these two positions. The toeless foot print is then adjusted to the perpendicular line. Finally, the length of toeless foot print and the footprint arch index values were calculated by applying Euclidean method to this adjusted toeless foot print. Footprint arch types are then classified by using the cutoff value of arch index. In this experiment, the average accuracy for automatic computing arch index values in the footprint image is 93.99% and the accuracy for classification of the footprint arch type is 84.40% in comparison with the manual computation from the expert. © 2017 IEEE.