Face detection with Euler number algorithm based on morphological operators

Abstract

This work conducts a practical study devoted to the Euler number utilization in face detection from binary and morphological processed images. Binary images are affected by noise and texture variability. Morphological operations change and reorder the pixel without acting on their values but generate a new image by either stripping away a layer of pixels or adding a layer of pixels to the boundaries of region of interest. The proposed algorithm binarized the images and then morphologically maps these images to establish the efficacy of different morphological operators. We investigated the performance of the algorithm on a database consisting of 95 face images rotated in the range of 0° to 90°. A similarity comparison study is performed by using an image quality metric called structural similarity index.