Detection & Classification of Lung Cancer at an Early Stage by Applying Feature Extraction Optimization and Neural Network on Hybrid Structure
Pankaj Nanglia1, Aprana N Mahajan2, Paramjit Singh3, Davinder Rathee4
1Pankaj Nanglia*, Currently Pursuing Ph.D in Electronics and Communication Engineering in Maharaja Agrasen University, Himachal Pradesh, India.
2Dr Aprana N Mahajan is Professor in Electronics and Communication Engineering in Maharaja Agrasen University, Himachal Pradesh, India.
3Paramjit Singh, Currently Pursuing Ph.D in Electrical & Electronics Engineering in Maharaja Agrasen University, Himachal Pradesh, India.
4Dr Davinder Rathee, Associate Professor in Electronics and Communication Engineering in Maharaja Agrasen University, Himachal Pradesh, India,.
Manuscript received on November 19, 2019. | Revised Manuscript received on 25 November, 2019. | Manuscript published on December 10, 2019. | PP: 3737-3745 | Volume-9 Issue-2, December 2019. | Retrieval Number: B6220129219/2019©BEIESP | DOI: 10.35940/ijitee.B6220.129219
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: As of now the detection and classification of lung cancer disease is one of the most tedious tasks in the field of medical area. In the diversified sector of medical industry usage of technology plays a very important role. Detection and diagnosis of the lung cancer at an early stage with more accuracy is the most challenging task. So, in this research article 400 set of images has been used for this experiment. Best feature extraction technique and best feature optimization technique has been analyzed on the basis of parameter minimum execution time with minimum error rate. Then finest selection of features leads to an optimal classification. In this context, one of the best classification algorithm the support vector machine has been proposed in this hybrid model for the binary classification. Further Feed forward back propagation neural network has been implemented with SVM. This proposed hybrid model reduces the complexity of the system on the basis of minimum execution time that is 1.94 sec. with minimum error rate 29.25. Further better classification accuracy 99.6507% has been achieved by using this unique hybrid model.
Keywords: Hybrid Structure, Lung Cancer Detection, Feature Extraction, SIFT, SURF, PCA and Feature optimization.
Scope of the Article: Classification