Multiple object detection in images using template matching

Abstract

Template matching is an application in computer vision for finding similar objects between the images. Template matching is a key component in image analysis process for mapping similar patterns between images. The detection and matching of objects has gained widespread application in quality control, manufacturing and medical applications. The major challenges in template matching process are occlusion, background noise and non-rigid transformation particularly in medical images. The present work proposes a template matching method based on correlation analysis techniques to detect multiple objects occurrence in an image. The method has been experimented for various real-time applications. In medical images, the proposed method gave an accuracy of 86% in detecting the nodules in lung CT images.