Effect of Changing Distances for Extracting Image Information for Error Reduction of Mouth Features

Abstract

Verity of information of the image in the form of image features is used in different application like pattern recognition, feature matching, image segmentation, image fusion, video processing, visual surveillance, medical diagnosis, traffic safety monitoring, remote sensing, human computer interaction, etc. Image is defined precisely and uniquely with the help of features which are useful in classifying and recognition of images. Extracting information from features of the image is a complex and diverse phenomenon. Retrieval of correct image information becomes very difficult for error reduction in the image processing. Lighting effect, zoom, distance, position, color model selected, angle of the object from camera etc. are considerable factors that affects the accuracy of feature \ndetection from the image. In this paper we studied and experimentation using Viola Jones algorithm are performed on distance and zoom for object detection for mouth feature detection on primary face database which is captured by smartphone. Analysis of the result concludes that as distance between object and camera increases, false negatives (Type II error) increase in mouth feature detection and it goes increasingly if the camera goes far away from object. These false negatives can be reduced by increasing zoom of the camera to achieve the accuracy and improve the mouth feature detection.