Articles

A Big Data Based Edge Detection Method for Image Pattern Recognition - A Survey

S. Dhivya, Dr.R. Shanmugavadivu,
Article Date Published : 22 March 2018 | Page No.: 23755-23760 | Google Scholar

Downloads

Download data is not yet available.

Abstract

In Today’s era Big Data is one of the most well-known research area that try to solve many research problems. The focus is mainly on how to come out those problems of Big Data and it could be handling in recent systems. Image mining and genetic algorithm is used to automate the process of images, patterns, data sets and etc. Image mining is used to extract the hidden images from the set of images. Genetic algorithm is also quite effective in solving certain optimization and intelligence problems and it is used in many applications, including image pattern recognition. The survey paper reviews of Big Data with edge detection methods on various types of images. In edge detection image pattern recognition is to choose the best images from the group of images by using both image mining and genetic algorithm techniques

1. Introduction:

Big Data is used both structured and unstructured data that is so large it is difficult to process using traditional database and software techniques. A data is classified as Big Data if it effectively satisfies any one of the requirements of Big Data, namely; Volume, Velocity and Variety. On the other hand, Big Data also arises with many challenges, such as difficulties in data capture, data storage, data analysis and data visualizations. Big data sets are used for minimizing risk, identifying unfamiliar objects and uncover hidden patterns. Data mining and Big Data are related to each other. Both containing more number of data’s and the data’s are extracted from the different data sets. Here, we have taken Big Data as an image. The clustering technique is used for grouping the images. After clustering the images, applies edge detection based genetic algorithm techniques like selection, crossover and mutation. In edge detection images are converted into gray scale images. For extracting the images, mining techniques will be used. Image mining is used to extract the images patterns from the large set of images, which draws the basic principles from concepts in databases, machine learning, statistics, pattern recognition and 'soft' computing. In image processing, edge detection is the one of the image processing technique for identifying the objects within the images.

There are many techniques used in edge detection but genetic algorithm is determined by a fitness function define to evaluate a solution’s ability to deal with a given task, ending up in bringing out a best possible solution. These paper discuses various techniques used for Edge Detection based Genetic Algorithms.

2. Literature Review

Various algorithms and techniques are used for agricultural image processing method. Here, we studied various plant Phenotyping image segmentation methods. Pascal Fua and Yvan G. Leclerc, describes standard edge detection method and it fails to find most of the relavent edges so author proposed the new method for images in initial segmentation [ 9 ]. Massimo Minervini , Andreas Fischbach , Hanno Scharr , Sotirios A. Tsaftaris, describes a collection of benchmark datasets of raw and annotated top-view color images of rosette plants which describe plant material, imaging setup and procedures for different experiments: [ 10 ]. Eren Erdal Aksoy, Alexey Abramov, Florentin Wörgötter, Hanno Scharr, Andreas Fischbach, Babette Dellen, preprocessing, leaf segmentation, and leaf tracking. Leaf-shape models are applied to improve leaf segmentation, and further used for measuring leaf sizes and handling occlusions [ 11 ]. J.Jeya Priyankha and K.Suresh kumar, describes the preprocessed image is subjected to K means clustering to get infected part of the leaf which is subjected to morphological processing to expanding the infected area. The author uses the HOG and SVM classifier algorithms[ 12 ]. Arti Singh, Baskar Ganapathysubramanian, Asheesh Kumar Singh, Soumik sarkar, describes the various methods about plant phenotyping and how the leaf images are extracted by using various machine learning and data mining techniques[ 13 ]. led Bhanu et al. to adopt a GA to determine the parameter set that optimise the output of an existing segmentation algorithm under various conditions of image acquisition and is namely Phoenix segmentation algorithm[ 18 ]. Another situation wherein GAs may be useful tools is illustrated by the work of Yoshimura and Oe [ 21 ]. For example, Bhandarkar et al. defined a multiterm cost function, which is minimized using a GAevolved edge configuration. The idea was to solve medical image problems, namely edgedetection. In their approach to image segmentation, edge detection is cast as the problem of minimising an objective cost function over the space of all possible edge configurations and a population of edge images is evolved using specialised operators. Fuzzy GA fitness functions were also considered by Chun and Yang, mapping a regionbased segmentation onto the binary string representing an individual, and evolving a population of possible segmentations[ 18 ]. Other GA approaches for image segmentation include manually-traced contours by Cagnoni et al., methods by Andrey, artificial ant colonies by Ramos, Koza’s genetic programming paradigm, Poli’s GP work, etc.,

3. Data Mining

Data Mining plays the very important role in the analysis of data. Data Mining is used to extract the hidden data from the large set of database. It is define as a “process of discovering new significant correlation, patterns into massive amount of information stored in warehouse [ 2 ]. Data Mining is also known as “Knowledge Discovery from data” (KDD) which consists of the following steps data selection, data cleaning, data integration, data transformation, pattern evaluation and knowledge presentation [ 5 ]. Knowledge discovery has been applied in many fields like health care and insurance, medicine, agricultural, assessment of credit card applications and so on. The following figure shows the process of both knowledge discoveries from Fig1.

Figure 1.Distribution in groups of patients by the stroke type.

4. Clustering

Clustering is the unsupervised method and there is no target field which also uses the bottomup approach. Clustering is a method for grouping the objects based on its attributes found in the data describing the objects or their relationships. A cluster is a collection of objects in a group which are “similar” to one other and different from the objects in other clusters. An image can be segmented based on its keyword (metadata) or its content (description)[ 8 ][ 7 ]. The clustering uses the various methods like partitioning, hierarchical, density-based, grid-based, model-based and constraint-based methods and Clustering algorithms can be classified into two main categories Linear clustering algorithms and Nonlinear clustering algorithms.

5. Image Edge Detection

Edge detection is the method of identifying edges in the images. Edge detection is used for image segmentation and data extraction in areas such as image processing, computer vision, and machine vision. The use of edge detection method is to reduce unnecessary information in the images and also extract the significant characteristics of the images such as corners, lines and curves. Edge generally occurs between objects and backgrounds, objects and objects, primitives and primitives. Edge is a boundary between two homogeneous regions.

In image processing, most of the shape information of an image is enclosed in edges. So, first we detect these edges in images by using gradient and zero-crossing method and then enhancing those detected areas of image which contains edges, sharpness of the image will increase and image will become clearer. The basic idea of edge detection is as follows: First, use edge enhancement operator to highlight the local edge of the image. Then, define the pixel "edge strength" and set the threshold to extract the edge point set. However, because of the noise and the blurring image, the edge detected may not be continuous [ 3 ]. The following flow chart describes the steps for image processing in edge detection Fig2.

Figure 1.Distribution in groups of patients by the stroke type.

6. Edge Detection Methods

The edge detection uses Gradient and zerocrossing methods.

Gradient Method

Prewitt Operator

Sobel Operator

Canny Edge Detection

Roberst Edge Detection

3. Non-maximum suppression: maxima should be marked as edges. Only local

4. Double thresholding: Potential edges are determined by thresholding. 5. Edge tracking by hysteresis: Final edges are determined by suppressing all edges that are not connected to a very certain (strong) edge.

Prewitt Operator

Prewitt operator is used for detecting edges horizontally and vertically. It is limited to 8 possible directions; however knowledge shows that most direct direction estimates are not much more perfect. In Fig3, This gradient based edge detector is estimated in the 3x3 neighborhood for eight directions.

Figure 1.Distribution in groups of patients by the stroke type.

Sobel Operator

Sobel edge detection detects all edges of the images which also has some advantage of providing both a differencing and smoothing effect. It is implemented as the sum of two directional edge enhancement operations. The operator consists of a pair of 3×3 convolution kernels as shown in Fig4. The sobel operator is very similar to Prewitt operator. It is also a derivate mask and is used for edge detection. It also calculates edges in both horizontal and vertical direction.

Figure 1.Distribution in groups of patients by the stroke type.

Canny Edge Detection

Prewitt Operator

Sobel Operator

Canny Edge Detection

Roberst Edge Detection

3. Non-maximum suppression: maxima should be marked as edges. Only local

4. Double thresholding: Potential edges are determined by thresholding. 5. Edge tracking by hysteresis: Final edges are determined by suppressing all edges that are not connected to a very certain (strong) edge.

Roberts Edge Detection

In fig5, It performs a simple, quick to compute, 2-D spatial gradient measurement on an image. This method emphasizes regions of high spatial frequency which often correspond to edges. The input to the operator is a greyscale image the same as to the output is the most common usage for this technique. This is very similar to sobel operator.

Figure 1.Distribution in groups of patients by the stroke type.

Zero-crossing Method

 Laplacian Operator.

Laplacian Operator

Laplacian Operator is also a derivative operator which is used to find edges in an image. Laplacian is a second order derivative mask. It can be further divided into positive laplacian and negative laplacian. The Laplacian L(x,y) of an image with pixel intensity values I(x,y) is given by:

Figure 1.Distribution in groups of patients by the stroke type.

All these masks find edges. Some find horizontally and vertically, some find in one direction only and some find in all the directions.

7. Edge Dtection Based Genetic Algorithm

Genetic algorithms are excellence for searching through large and complex data sets. They also considered capable of finding reasonable solution to complex issues as they are highly capable of solving unconstraint and constraint optimization issues. The GA uses the tree main types of rules at each step to create the next generation from the current population i.e Selection rules, Crossover and Mutation. The selection evaluates each individual and keeps only the fittest ones in the population [ 19 ]. In addition to those fittest individuals, some less fit ones could be selected according to a small probability [ 20 ]. The others are removed from the current population. The crossover recombines two individuals to have new ones which might be better. The mutation operator induces changes in a small number of chromosomes units. Its purpose Page 27758 is to maintain the population diversified enough during the optimization process [ 18 ]. One reason for using GA is to deal large, complex search spaces where only minimum knowledge is available about the objective function

8. Conclusion

In this paper, studied the most commonly used edge detection techniques of Gradient-based and Laplacian based Edge Detection. Choosing a suitable method for edge detection is based on the some environmental conditions. Each technique has its own advantages and disadvantages. This review paper will be helpful for the researchers in understanding the concept of edge detection.

References

  1. Kumar Kamlesh, Li Jian-Ping, Khan SaeedAhmed. Comparative Study on Various Edge Detection Techniques for 2-D Image 2015;9(1):6-10. CrossRef | Google Scholar
  2. Nagpal Anushree, Tajdar andAtiqua, Siddiqui MasoodAhsan, Hassan Mohammad, Gaur Suman, Hashim Mohammad, and and and. Beyond the Brothels: A Qualitative Study of the Destitute Female Sex Workers of Central Delhi Red Light Area 2017;9(1):26-36. CrossRef | Google Scholar
  3. Gao Wenshuo, Zhang Xiaoguang, Yang Lei, Liu Huizhong. An improved Sobel edge detection 2010. CrossRef | Google Scholar
  4. Ye Hanmin, Yan Shili, Huang Peiliang. 2D Otsu image segmentation based on cellular genetic algorithm 2017. CrossRef | Google Scholar
  5. Cios KrzysztofJ, Kurgan LukaszA. Trends in Data Mining and Knowledge Discovery ;:1-26. CrossRef | Google Scholar
  6. Sharma Priyansh, Suji Jenkin. A Review on Image Segmentation with its Clustering Techniques 2013-may;:209-218. CrossRef | Google Scholar
  7. kaur Jaskirat. Performance analysis of clustering based image segmentation and optimization methods 2012. CrossRef | Google Scholar
  8. kaur Jaskirat. Performance analysis of clustering based image segmentation and optimization methods 2012. CrossRef | Google Scholar
  9. Fua Pascal, Leclerc YvanG. Model driven edge detection 1990-dec;:45-56. CrossRef | Google Scholar
  10. Minervini Massimo, Fischbach Andreas, Scharr Hanno, Tsaftaris SotiriosA. Finely-grained annotated datasets for image-based plant phenotyping 2016-oct;:80-89. CrossRef | Google Scholar
  11. Aksoy ErenErdal, Abramov Alexey, WĂśrgĂśtter Florentin, Scharr Hanno, Fischbach Andreas, Dellen Babette. Modeling leaf growth of rosette plants using infrared stereo image sequences 2015-jan;:78-90. CrossRef | Google Scholar
  12. Singh Arti, Ganapathysubramanian Baskar, Singh AsheeshKumar, Sarkar Soumik. Machine Learning for High-Throughput Stress Phenotyping in Plants 2016-feb;:110-124. CrossRef | Google Scholar
  13. Deshmukh Jyoti, Bhosle Udhav. Image Mining Using Association Rule for Medical Image Dataset 2016;:117-124. CrossRef | Google Scholar
  14. Muthukrishnan R, Radha M. Edge Detection Techniques For Image Segmentation 2011-dec;:259-267. CrossRef | Google Scholar
  15. Patil YogeshUttam. Quantitative Review Techniques of Edge Detection Operators. 2017-mar;:24-26. CrossRef | Google Scholar
  16. Lee MK, Leung SW, Pun TL, Cheung HL, Lee AMK. Edge detection by genetic algorithm . CrossRef | Google Scholar
  17. Castillejos Heydy, Ponomaryov Volodymyr. Fuzzy Image Segmentation Algorithms in Wavelet Domain 2012. CrossRef | Google Scholar
  18. Borji A, Hamidi M, moghadam AMEftekhari. CLPSO-based Fuzzy Color Image Segmentation 2007. CrossRef | Google Scholar
  19. Wen Xian-Bin, Zhang Hua, Jiang Ze-Tao. Multiscale Unsupervised Segmentation of SAR Imagery Using the Genetic Algorithm 2008-mar;:1704-1711. CrossRef | Google Scholar
  20. Rangoussi M, Giannakis GB. FIR modeling using log-bispectra: weighted least-squares algorithms and performance analysis . CrossRef | Google Scholar

Author's Affiliation

Copyrights & License

International Journal of Engineering and Computer Science, 2018.
Creative Commons License

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

Article Details


Issue: Vol 7 No 03 (2018)
Page No.: 23755-23760
Section: Articles
DOI:

How to Cite

Dhivya, S., & Shanmugavadivu, D. (2018). A Big Data Based Edge Detection Method for Image Pattern Recognition - A Survey. International Journal of Engineering and Computer Science, 7(03), 23755-23760. Retrieved from http://ijecs.in/index.php/ijecs/article/view/3993

Download Citation

  • HTML Viewed - 1070 Times
  • PDF Downloaded - 144 Times
  • XML Downloaded - 0 Times