COLOR IMAGE SEGMENTATION USING DISCRETE WAVELET TRANSFORM AND IMPROVED SALIENCY MAP

Abstract
The main objective of this paper is to apply Self-Organizing Map (SOM) based Enhanced Fuzzy C-Means (EFCM) algorithm to color image segmentation which is a crucial problem in computer vision and pattern recognition. Discrete Wavelet Transform (DWT) is applied on the color image. The decomposition level is set to 3. Approximation coefficients are extracted from the decomposed image. Improved Saliency Map (ISM) is computed. Self-Organizing Map (SOM) is trained with approximation coefficients along with ISM values. The resultant image is clustered with EFCM algorithm. Proposed method is validated on Berkeley segmentation dataset and other natural color images. Performance of the method is evaluated by using accuracy, precision, recall, entropy and time. Simulation results showed that the proposed method can achieve good segmentation results with low computational complexity than other methods considered for comparison from the literature.

Authors
M Sivasubramanian1, P Kumar2, M Sivajothi3
Manonmaniam Sundaranar University, India1,2, Sri Parasakthi College for Women, India3

Keywords
Color Image Segmentation, Discrete Wavelet Transform, Improved Saliency Map and SOM Enhance Fuzzy C-means Clustering
Published By :
ICTACT
Published In :
ICTACT Journal on Image and Video Processing
( Volume: 9 , Issue: 4 )
Date of Publication :
May 2019

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.