Digital Image Processing

Lydia Liz Lukose ¹, Niya Joseph ¹, Kethsy Prabavathy ²

Affiliations
1 Karunya University, IN
2 Department of Computer Science at Karunya University, Coimbatore, IN

Abstract

Here we introduce a novel approach for multiphase image segmentation. The approach presents a unified framework that unifies three basic segmentation approaches; level set methods, graph cut algorithms and unsupervised multiphase segmentation. Here we have presented an image segmentation approach that have the advantages of segmenting the image into multiple phases where the favorable number of segments the image to be segmented will be automatically chooses by the algorithm itself. The unsupervised multiphase segmentation will provide the separation of image into favorable number of multiple phases. Our main objective is to segment the image robust to noise, blurred edges and topology changes and to achieve global optimization and speed. Here we are segmenting the image by constructing graph on the basis of the intensity values of each pixel in the image. The graph cuts method will provide robustness towards noise and the brute force numerical algorithm for unsupervised multiphase segmentation will provide fast and efficient segmentation and global optimization.

Keywords

Phase, Balance, Graph Cuts.

Full Text

     

Niya Joseph ¹, Lydia Liz Lukose ¹, Kethsy Prabavathy ²

Affiliations
1 Karunya University, IN
2 Department of Computer Science at Karunya University, Coimbatore, IN

Abstract

For retrieving images from the database, content based image retrieval is a widely used technique. In order to improve the performance of content based image retrieval, relevance feedback has been widely used. Here in this method we proposed a technique called biased discriminant Euclidean algorithm, in which it considers both intraclass geometry and interclass discrimination. Labeled and unlabeled samples can be used in order to get the relevant images. To enhance the relevance feedback performance for the unlabelled samples we integrate a concept called Semi-Supervised Orthogonal Discriminant Analysis via Label Propagation with semi-supervised biased discriminant Euclidean embedding (SBDEE). It propagates label information from labeled data to unlabeled data according to the distribution of labeled and unlabeled data. Thus the distribution of the unlabeled data can be effectively explored to learn a better subspace. The scatter matrices based on soft label learned by label propagation are defined to perform the discriminant analysis, which gives us a general framework to extend many variants of supervised discriminant analysis to the semi-supervised ones.

Keywords

Content Based Image Retrieval, Dimensionality Reduction, Discriminant Analysis Feature Extraction, Semi-Supervised Learning, Similarity Measure.

Full Text