Open AccessColour Image Segmentation using Background Subtraction with Global and Local Threshold
Author(s) -
Amanpreet Kaur
Publication year - 2021
Publication title -
international journal for research in applied science and engineering technology
Language(s) - English
Resource type - Journals
ISSN - 2321-9653
DOI - 10.22214/ijraset.2021.35340
Subject(s) - artificial intelligence , computer vision , computer science , image segmentation , thresholding , scale space segmentation , segmentation based object categorization , segmentation , digital image processing , digital image , region growing , background subtraction , image processing , image texture , pattern recognition (psychology) , minimum spanning tree based segmentation , range segmentation , image (mathematics) , pixel
Image segmentation is one of the fundamental and essential steps in all the major applications of digital image processing. In this process the digital image is divided into various regions which are also known as segments. These segmented parts of the digital image could be used for further processing like detection of types of objects present in the segmented region, various tumors present in the digital images or the scene understanding process. Usually segmentation is classified as local segmentation and the global segmentation. Image segmentation is also classified on the basis of digital image properties also. In this case it is of two types. First one is non continuity detection and second one is the continuous detection. Various image segmentation techniques are proposed by researchers which have various limitations. Some techniques do not split the region uniformly and other techniques take enough time and memory for the processing of digital image. In this research work both the local and global thresholding concept is used to get the segmented regions of the image. The proposed technique will be able to extract the segmented objects from the digital image. To check the authenticity and efficiency of the proposed technique, it will be compared with other well known techniques of image segmentation using background subtraction of colored digital images. Time of computation, sensitivity and accuracy are used as objective parameters for the performance evaluation of the techniques. For the subjective evaluation visual quality of the digital image is used for performance evaluation.