JPM1410 Images as Occlusions of Textures: A Framework for Segmentation
- 1. Images as Occlusions of Textures: A Framework for Segmentation ABSTRACT: We propose a new mathematical and algorithmic framework for unsupervised image segmentation, which is a critical step in a wide variety of image processing applications. We have found that most existing segmentation methods are not successful on histopathology images, which prompted us to investigate segmentation of a broader class of images, namely those without clear edges between the regions to be segmented. We model these images as occlusions of random images, which we call textures, and show that local histograms are a useful tool for segmenting them. Based on our theoretical results, we describe a flexible segmentation framework that draws on existing work on nonnegative matrix factorization and image deconvolution. Results on synthetic texture mosaics and real histology images show the promise of the method. EXISTING SYSTEM: In an Existing System, a wide variety of classic approaches to generic image segmentation, including graph cuts , active contours, level sets, Gabor filtering and
- 2. clustering, random fields, watersheds, region growing , and mean shift . In addition, there also exist engineered segmentation systems, such as BlobWorld, JSEG, EDISON, and CTex. Broadly, these methods vary in how the segmentation regions are parametrized and whether edge or region information is used. Each existing method makes assumptions about the images it aims to segment. When these assumptions are met, the method works, and when they are not, it fails. Because many of these assumptions are implicit, selecting a method to use on a new segmentation problem involves educated guessing. When no suitable method can be found, a new method is designed. DISADVANTAGES OF EXISTING SYSTEM: Even with this plethora of methods to try, working on a new segmentation problem is not trivial. Each existing method makes assumptions about the images it aims to segment. When these assumptions are met, the method works, and when they are not, it fails. Because many of these assumptions are implicit, selecting a method to use on a new segmentation problem involves educated guessing. Our working hypothesis is that this task is difficult because tissues are complicated, while tissue boundaries are sometimes subtle and not marked by edges.
- 3. PROPOSED SYSTEM: We propose a mathematical framework for image segmentation which models images as occlusions of textures. Given an image formed according to this model, we prove that its local value histograms will approximately be convex combinations of the value distributions of its component textures. Based on this result, we present a new algorithmic framework for image segmentation based on histogram factorization and deconvolution. ADVANTAGES OF PROPOSED SYSTEM: Given an image formed according to this model, we prove that its local value histograms will approximately be convex combinations of the value distributions of its component textures. Based on this theorem, we proposed a segmentation framework that first discovers the value distributions of the component textures of an image, finds the contribution of each texture to every local histogram of the image, then deconvolves the contributions to recover a segmentation.
- 4. SYSTEM ARCHITECTURE: SYSTEM REQUIREMENTS: HARDWARE REQUIREMENTS: System : Pentium IV 2.4 GHz. Hard Disk : 40 GB. Floppy Drive : 1.44 Mb. Monitor : 15 VGA Colour. Mouse : Logitech. Ram : 512 Mb. SOFTWARE REQUIREMENTS: Operating system : Windows XP/7.
- 5. Coding Language : MATLAB Tool : MATLAB R 2007B REFERENCE: QYong Lin, Eliza Yingzi Du, Senior Member, IEEE, Zhi Zhou, Student Member, IEEE, and N. Luke Thomas ,“An Efficient Parallel Approach for Sclera Vein Recognition”, IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, VOL. 9, NO. 2, FEBRUARY 2014.