International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395 -0056
Volume: 04 Issue: 05 | May -2017
p-ISSN: 2395-0072
www.irjet.net
Reclassification of low Intensity pixels using seed growing V.Saran Raj1, S.Santhosh Rewanth2, G.Kaliyugavaratharaj3, M.Poonkodi,B.E., M.Tech4 1Student,
Dept. of Computer Science Engineering SRM University, Tamilnadu, India Dept. of Computer Science Engineering SRM University, Tamilnadu, India 3Student, Dept. of Computer Science Engineering SRM University, Tamilnadu, India 4 Assistant Professor, Dept. of Computer Science Engineering SRM University, Tamilnadu, India 2Student,
---------------------------------------------------------------------***--------------------------------------------------------------------1.1 Blurring Abstract - Our objective is to reclassify low intensity pixels from seed chosen by random walker and seed growing methods to grow seed. The proposed system, which is an enhanced algorithm used with Local Binary Pattern, is to process the same image so that it can be proved that the enhanced method gives more clarity to the image than the random walker and seed growing.
Blurring is present in all imaging processes including vision, photography, and medical imaging methods. It is important to understand the smallest details of the image which depends on the amount of blur produced by the imaging procedure. The amount of blurring can be measured as the dimension of the blurred image of a very small object . The blur also has different shape that depends on the source of blur. For example, some x-ray system produce round blur patterns.
Key Words: Image segmentation, Radom Walker, Seed Growing, Local Binary Pattern
1. INTRODUCTION
The intensity distribution in the blur area is also an important factor in addition to blur size and shape. One of the distribution patterns is high intensity near the center with a gradual reduction of intensity toward the periphery. The blur limits the amount of details that can be extracted from the objects. This will result in spreads of small objects into surrounding background area.
The image processing can take many forms: filtering, compression, feature extraction and enhancement. Image segmentation is the first step in image processing to distinguish the object and background. Through image segmentation, the image is transformed into various phases, however the process tracks the important features of each phase. Among various segmentation technologies available, choosing an appropriate technology is decided by the particular type of image and the characteristics of the problem being considered. There is no universally accepted method for image segmentations. The techniques are combined with the domain knowledge to solve an image segmentation problem effectively. The image segmentation techniques are classified into two categories: Region based image segmentation and Edge based image segmentation.
The visibility of an object depends on the relationship between object size and blur value. If the blur value is comparatively very less than the dimension of the object, the visibility will not be affected much. The image with much details and distinct boundaries is described as sharp. The presence of blur produces unsharpness. Un-sharpness is especially notable at the boundaries and edges. It is very important to the technique to separate objects that are close together. To resolve the segmentation of the objects, their separation distance must be increased in proportion to the amount of blur present.
Random walk, graph cut and grab cut are few of the image segmentation techniques. Each technique has its merits and demerits. In this paper, we analyze how the local binary pattern technique is used to enhance the intensity of blurred image pixels and compare its performance with other methods.
Š 2017, IRJET
|
Impact Factor value: 5.181
It is important to identify the high intensity region and low intensity region. High intensity region will tend to lose more signal to their surrounding pixels than they gain, and
|
ISO 9001:2008 Certified Journal
|
Page 180