Indian Journal of Science and Technology
Year: -0001, Volume: 9, Issue: 1, Pages: 1-12
M. Mercy Theresa1* and V. Subbiah Bharathi2
1Faculty of Electronics and Communication Engineering, Sathyabama University, Chennai – 600119, Tamil Nadu, India; [email protected] 2S. R. M. University, Chennai – 603203, Tamil Nadu, India; [email protected]
*Author For Correspondence
M. Mercy Theresa
Faculty of Electronics and Communication Engineering, Sathyabama University, Chennai – 600119, Tamil Nadu, India; [email protected]
Background and Objectives: The main objective of this paper is to extract the lung nodules in Chest Radiography (CR) for identifying and locating the presence of small lesion or nodules. Methods and Statistical Analysis: Four stages of works are carried out such as: 1). Image Registration using geometrical transformation method, 2). Lung Segmentation using thresholding method, 3). Feature Extraction using Complex Wavelet Transformation method and Shearlet transformation method and finally 4). Image Classification using Random Forest method and classify the images are normal or abnormal. To confirm the abnormality of the segmented output the results are compared with the labeled ground truth images given by medical experts. Findings:The entire proposed approach is implemented in MATLAB software and the performance is verified by comparing the results obtained from CWT results with the ST results. The proposed experiment is carried out on JSRT dataset and obtained the accuracy 96% which is better than the existing approach’s obtained accuracy is 95.4%. Conclusion and Improvements: The proposed approach in this paper is better than the existing approaches for lung nodule detection and identification through segmentation in chest radiography. In future work the performance of the proposed work is evaluated by comparing the obtained results with the other existing transformation methods.
Keywords: Chest Radiography, Complex Wavelet Transformation, Lung Nodule Detection, Shearlet Transformation, CAD
Subscribe now for latest articles and news.