Face and Human Detection using Gradient Pattern with Nearest Neighbor Technique

Abstract

Detecting and classifying the image is an important topic of computer technology that determines the sizes in digital images. It is mainly used for authentication and identification purpose. Intensity transform features convert the intensity (pixel color) into an encoded value by comparing the pixel. Normally, detection methods take the pixel color directly as information cues. These schemes are more sensitive to noise, variations in pose and changes in illumination. To tackle these effects by using GP (Gradient pattern) and nearest neighbor algorithm. Gradient feature makes the variation of edges. Nearest neighbor reduces the variations of misclassified area. The images are classified by using Background variations. The proposed approach is to rectify the intensity differences at various locations, provides a minimum processing time and the misdetections are prevented in crowd environment.

Keywords—Gradient feature pattern (GFP) and K nearest neighbor (KNN).