Telecommunications | Journals | Books and Articles

The IUP Journal of Telecommunications

Feature Priors for Image Segmentation Using HMRF Algorithm

Article Details

Pub. Date	: Jan' 2021
Product Name	: The IUP Journal of Telecommunications
Product Type	: Article
Product Code	: IJTC40221
Author Name	: Gyanender Kumar and Lincoln Hadda
Availability	: YES
Subject/Domain	: Arts & Humanities
Download Format	: PDF Format
No. of Pages	: 13

Price

Download

Abstract

Image segmentation is a process of dividing the image into some distinct regions. These regions are specially coherent in nature and have similar attributes. This technique is widely used for image analysis and to interpret the desired feature. The paper studies the Hidden Markov Random Fields (HMRF) and finds its Expectation Maximization (EM) algorithms. The main idea behind developing HMRF is to adjoin the "data faithfulness" and "model smoothness" which has very similar nature with the active contours, Gradient Vector Flow (GVF), graph cuts and random walks. The paper also uses HMRF-EM along with the Gaussian mixture models, and then color image segmentation process. These algorithms are implemented in MATLAB. In color image segmentation experiments, it is observed that the results obtained from HMRF segmentation are much smoother than the direct k-means clustering. The segmented object is much closer to the original shape than clustering. The segmentation time for Bacteria 1, Bacteria 2, SAR and brain images is 0.35, 0.43, 0.12 and 0.12, respectively. The accuracy for Bacteria 1, Bacteria 2, SAR and brain images is 97.70%, 98.06%, 98.89% and 97.35%, respectively.

Introduction

If we study the image processing system, the resultant image may contain some irregularities or defects that may affect our process. Furthermore, these kinds of defects can be adjusted by various kinds of techniques such as increasing the number of pictures from the same scene which decreases the effect of defect and using higher quality instruments, but such methods which are based on the external hardware consume more time and they increase the cost too (Manisha and Geetanjali, 2018). So to avoid such effect of external hardware, we often use computer programs which consume very less time and reduce the cost (Anuva et al., 2015). For example, to remove the noise defect, we can use smooth filter which effectively reduces the noise content and filters the image, or to change the contrast level in a low contrast image,