Deakin University
Browse
samadiani-usingmorphological-2015.pdf (2.53 MB)

Using morphological transforms to enhance the contrast of medical images

Download (2.53 MB)
journal contribution
posted on 2015-06-01, 00:00 authored by H Hassanpour, Najmeh Samadiani, S M Mahdi Salehi
Medical imaging plays an important role in monitoring the patient's health condition and providing an effective treatment. However, the existence of several objects overlapping in an image and the close proximity of adjacent pixels values in medical images make the diagnostic process a difficult task. To cope with such problems, this paper presents a new method based on morphological transforms to enhance the quality of various medical images. In this method, a disk-shaped mask whose size fits that of the original input image is chosen for morphological operations. Afterward, the proposed filter from the Top-Hat transforms is applied to the image, using the chosen mask in a multi-step process. At each step, the size of the mask is increased. Consequently, an enhanced image is provided for each mask size. The number of required steps and the final enhanced image are determined based on the Contrast Improvement Ratio (CIR) measure. Indeed, this approach applies an exfoliation process on the images, in which one or several objects in the image are prominently manifested using morphological filter, hence provide an appropriate image for analysis. The results in this research indicate that the proposed approach makes a better contrast and works much better than the other existing methods in improving the quality of medical images.

History

Journal

Egyptian journal of radiology and nuclear medicine

Volume

46

Issue

2

Pagination

481 - 489

Publisher

Elsevier

Location

Amsterdam, The Netherlands

ISSN

0378-603X

eISSN

2090-4762

Language

eng

Publication classification

C1.1 Refereed article in a scholarly journal

Usage metrics

    Research Publications

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC