Digital Image Processing

S. Deepa ¹, V. SubbiahBharathi ²

Affiliations
1 Sathyabama University, Chennai, IN
2 DMI College of Engineering, Chennai, IN

Abstract

In this paper we have proposed three different approaches for extracting the Region Of Interest (ROI) in digital mammogram images. Segmentation of the Region of Interest is the first and crucial step in the analysis of digital mammogram images since the success of any Computer Aided Diagnostic (CADx) system depends greatly on the accuracy of the segmentation of the ROI from the mammogram images. Finding an accurate, robust and efficient ROI segmentation technique still remains a challenging problem in digital mammography analysis. Three different segmentation techniques viz., Otsu‟s N thresholding method, k means clustering, and efficient marker controlled watershed transform are applied on the digital mammogram and their results are compared. The results are compared based on the ROI area extracted from the original mammogram image. From the results obtained Otsu‟s N thresholding and k means clustering methods prove to be promising compared to marker controlled watershed transform.

Keywords

Digital Mammograms, K Means Clustering, Marker Controlled Watershed Transform, Otsu‟s N Thresholding, Region of Interest.

Full Text

     

S. Deepa ¹, V. Sadasivam ²

Affiliations
1 Department of Information Technology, National College of Engineering, Tirunelveli, IN
2 PSN College of Engineering, Melathediyoor, Tirunelveli, IN

Abstract

Multispectral images are images with high spatial, spectral and radiometric resolution. Efficient multispectral image compression plays a key role in most of the geographical applications. The three important phases involved in a adaptive technique are transformation, clustering and encoding based on the resultant clusters. The existing methods like SPIHT, STW, WDR, and ASWDR are used as the base for Adaptive encoding techniques. The methods adopted are the recent wavelet compression techniques adopted for high dimensional images. The paper proposes the adaptive encoding technique and different combination of encoding techniques are evaluated and an efficient technique (E –Ad) is identified by comparing the results using the metrics Peak Signal to Noise Ratio as well as Compression Ratio. The proposed compression technique preserves the unique spectral characteristics of the multispectral image. The compressed image, can be efficiently used as preprocessing for many enhancement techniques and for object classification mechanisms. The E-Ad technique overtakes many of the state of art algorithms.

Keywords

ASWDR, E–Ad, SPIHT, STW, WDR.

Full Text