Mycobacterium Tuberculosis Bacilli cells Recognition using Multiple Classifier
Journal Title: International Journal of Research in Computer and Communication Technology - Year 2014, Vol 3, Issue 11
Abstract
Many computer science researchers have been working on developing and designing an automated system for fast & accurate recognition of tuberculosis (TB), which would ensure speed up of treatment process. Manually for proper identification of Mycobacterium tuberculosis require skilled person it is critical, time consuming and hectic too. Identification of TB bacilli manually sometime may confuse with some stain residue & non tuberculosis bacilli. Due to this we get the faulty results. For this reason need of atomization is required for exact identification of tuberculosis. We present an object, pattern (Mycobacterium Tuberculosis cell) recognition method using Support Vector Machine (SVM), k-nearest neighbor (Knn) and Gaussian membership function. The object i.e. M.TB bacilli cells are extracted using color segmentation method, from ZiehlNelson stained sputum smears images. These images have blue background on which red color M.TB bacilli cells. Images are by self developed data set, obtained using a bright field microscope. The various shapes of objects are extracted using segmentation method. The extracted shapes is the mixture of M. TB bacilli cells, non TB cells and few are stain residue etc. For proper recognition(M.TB cells) the shape based feature of these patterns are extracted on the basis of these futures data set the multiple classifiers like SVM, Knn and Gaussian are applied for recognition of M.TB bacilli cells. The recognition rate obtains using SVM is highest than Knn and Gaussian.
Authors and Affiliations
Jadhav Mukti E. , Kale K. V.
Keywords
Related Articles
- EP ID EP28111
- DOI -
- Views 289
- Downloads 0
Novel De-duplication Technique for Supporting Differential Privileges Check in H-Cloud
Clients are benefited storing their data in Cloud instead of local data management. But one critical challenge of cloud storage is management of huge volumes of data and lack of privacy and security which maximizes t...
Learning and Sequential Decision Making For Medical Data Streams Using Rl Algorithm
Data stream mining has obtained a high attraction due to the importance of its applications and increase in the generation of streaming information. The data streams are the set of data, which are moving in specified...
A Survey of Signcryption Its Type and Techniques
Signcryption is a combinational method of performing signatures of the encrypted data so that the data can be made secure and authenticated during the access of the data. Since there are various techniques are implem...
Low-Power Digital Signal Processor Architecture for Wireless Sensor Nodes
Radio communication exhibits the highest energy consumption in wireless sensor nodes. The sensor networks are intended to support a variety of applications . Providing sensor network with flexible nodes design will...
Medium Voltage High-Power Converter Topology for the Large Permanent Magnet Wind Generator Systems
This paper proposes a modular, medium voltage, highpower converter topology for the large permanent magnet wind generator system, eliminating the grid-side step-up transformer, which is desirable for both onshore and o...