S.P. Pavithrra, L. Sugumar, A. Vaishnavi, P. Vinoth

South Asian Journal ofEngineering and Technology Vol.4, No.1 (2018) 27–33

ISSN No: 2454-9614

[S1]

Classification of Lung Cancer using Image Processing Techniques

S.P. Pavithrra, L. Sugumar, A. Vaishnavi, P. Vinoth,

Mrs. Reeja Antony

Assitant professor,

Department of Biomedical Engineering

Velalar College Of Engineering And Technology.

.

*Corresponding Author:Mrs. Reeja Antony

Phone: +91-422-2692461;

E-mail: ,

Received: 18/1/2018, Revised: 25/2/2018 and Accepted: 18/3/2018

Abstract

Lung tumor extraction and analysis are challenging tasks in medical image processing. PET has become a useful medical diagnostic tool for the diagnosis of lung and other medical images. PET scan goes beyond anatomy to show what is happening at a cellular level in the tissue of the lung. A multistage system is used for lung tumor diagnosis and tumor region extraction. First, noise removal and histogram equalization is performed as the preprocessing step on the lung PT images. Statistical and texture features are extracted from these noise free lung PT images. Next phase of the proposed system is multiclass classification that is based on these extracted features. This method uses multiclass Support Vector Machine (SVM) to classify the tumor.

Keywords: Histogram equalization, Lung tumor, PET, Segmentation, Classification

1. Introduction

Humidity sensors have In nature, lung disease plays a major role in health issue. The breathing is affected first in any form of lung diseases. Some of the common lung disease are Acute bronchitis, asthma, Chronic Obstructive Pulmonary Disease (COPD), Acute Respiratory Distress Syndrome (ARDS) and Lung cancer. Lung cancer is a disease which involves the multiplication of abnormal cells which finally leads to tumor. Small cell carcinoma and non–small cell carcinoma are the two types of lung cancer. These types are used for treatment decisions and determining prognosis1. The only solution to cure lung cancer is to detect it at an early stage. Proper staging of lung cancer is important because the treatment options and prognosis differ significantly by stage. Mortality from lung cancer are expected to rise around 17 million worldwide in 2030. Early detection of lung cancer can increase the chance of survival among people. There are many techniques to diagnose the lung cancer, such as Chest Radiograph (X-ray), Computed Tomography (CT), Magnetic Resonance Imaging (MRI scan) and Sputum Cytology. However, most of these techniques are expensive and time consuming2. Therefore, there is a great need for a new technology to diagnose the lung cancer in its early stages. Image processing techniques provide a good quality tool for improving the manual analysis. Clustering has been a popular approach to untested pattern recognition. Image segmentation is important in the field of image understanding, image analysis, pattern recognition and computer vision. The principal goal of the segmentation process is to partition an image into regions that are homogeneous with respect to one or more characteristics and features. But, due to the impact of lighting in imaging process, sometimes the boundaries of object we get are not real, especially on the process of objects with varied topology structure, the traditional algorithms can’t determine the real boundaries. Objective of this study is to detect lung cancer using image processing techniques. PET scanned lung images of cancer patients are acquired from various databases. Using image processing techniques like pre-processing and feature extraction, area of interest is separated. Developing the algorithm, features are extracted from all the images. The parameter values obtained from these features are compared with the normal values suggested by a physician. From the comparison result, cancer stage is detected. This system can help in early detection of lung cancer more accurately3.

II LITERATURE REVIEW

All researchers have aim to develop a system which predict and detect the cancer in its early stages. They have also tried to improve the accuracy of the Early Prediction and Detection system by preprocessing, segmentation, feature extraction and classification techniques. The major contributions of the research are summarized below:

Priyanka Basak, Asoke Nath, (2017) have proposed that Gabor filter and watershed segmentation gives best results for pre-processing stage. After the extraction of interested region, three features are extracted i.e., area, perimeter and eccentricity for the classification of different stages of lung cancer4.

Aniket Gaikwad, Azharuddin Inamdar, Vikas Behera, (2016) have used 3 stages such as image segmentation, feature extraction stage, the watershed segmentation technique to improve more accurate result. They have also concluded that the technique gives more accuracy (84.55%) than other approach5.

David Ladron de Guevara Hernandez, (2015) have proposed that PET/CT imaging is the method of choice for the extra-encephalic staging of lung cancer patients with a better performance than the traditional imaging methods.

Dasu Vaman Ravi Prasad, (2013) proposed a method to improve the image quality and accuracy. They have used low pre-processing technique based on Gabor filter within Gaussian rules as the image quality and improvement depends on the enhancement stage. The enhanced region of the object of interest is then used as a basic foundation of feature extraction. Relying on general features, a normality comparison is made. In this research, the main detected features for accurate images comparison are pixels percentage and mask-labeling.

S Vishukumar K. Patela and Pavan Shrivastavab (2012) have proposed gabor filter for enhancement of medical images. In this paper authors mostly focus on significant improvement in contrast of masses along with the suppression of background tissues. It is obtained by tuning the parameters of the proposed transformation function in the specified range.

III METHODOLOGY

Image acquisition:

The PET scan image of a lung cancer patients is acquired. PEToffers substantialadvantages overanatomic imaging modalities in oncologic imaging.PETcan often distinguish between benign and malignant lesions whenCTand MRI cannot. The PET images are acquired

Image pre-processing:

This process is sub-divided into ‘Smoothing’ and ‘Enhancement’.

Smoothing: It controls the noise and further small instabilities in the image. Smoothing blurs out most of the edges that contain useful information. Most of the methods reduces noise by blurring the image irrespective of their implementation. Even a high resolution photo is bound to have some noise in it. For a high resolution picture a simple box blur may be sufficient. Mainly the idea of neighborhood filter is to calculate pixel weights, depending on their color’s similarity. We describe two methods: the use of ‘High pass filtering’ and ‘Median Filtering’7.

High pass filter: A high-pass filter (also known as a bass-cut filter) prohibits signals below a cutoff frequency (the stopband) and allows signals above the cutoff frequency (the passband). The output of this filter is directly proportional to rate of change of the input signal.

Median filter: It is a nonlinear digital filter. It is used to remove the noise in the image. To detect some edge in the image, firstly noise should be removed up to some threshold value and then edge removal is performed. Hence the median filter is placed before edge detector. Its main feature is to remove noise without affecting the edge8.

Enhancement: Enhancement is used to improve the perception of information in images for human viewers and to provide better input for other automated image processing techniques. The enhancement technique used in this proposed method is ‘Histogram equalisation’.

Histogram equalization: This method usually increases the globalcontrastof images, especially when the usabledataof the image is represented by close contrast values. Through this adjustment, theintensitiescan be better distributed on the histogram.

Segmentation:

Segmentation is the partation of an image. Segmentation is typically used to detect objects and boundaries of an image.

Otsu thresholding: It is the process of conversion of a grayscale image into a monochromatic one. Otsu's thresholding method involves iterating through all the possible threshold values and calculating a measure of spread for the pixel levels on each side of the threshold, i.e. the pixels that either fall in foreground or in background. The aim is to find the threshold value where the sum of foreground and background spreads at its minimum9.

K-means clustering: It is a type of unsupervised learning, which is used there is an unlabeled data (i.e., data without defined categories or groups). The goal of this algorithm is to find groups in the data, with the number of groups represented by the variableK. The algorithm works iteratively to assign each data point to one ofKgroups based on the features that are provided. Data points are clustered based on feature similarity. Rather than defining groups before looking at the data, clustering allows to find and analyze the groups that have formed organically10.

Fig. 2: SEGMENTED TUMOR

Feature extraction:

It is an important stage in the image processing technique. It detects desired portion or shape of an image. In the work, we have extracted Texture features, Statistical features and Structural Features for the segmented tumor 11.

Classification:

Support vector machines are supervised learning models that analyze data and recognize patterns, used for classification. The basic SVM takes a set of input data and for each given input, predicts which of two classes forms the input, making it a non-probabilistic binary linear classifier. SVM uses a kernel function which maps the given data into a different space and the separations can be made even with very complex boundaries12.

VI RESULTS AND CONCLUSION

The proposed method identifies the tumor with accuracy from the original image. K-means clustering gives the best results for the segmentation of tumor. From the extracted region of interest, statistical and texture features are extracted. These three features help to identify the stage of lung cancer using Support Vector Machine.

V. REFERENCES

[1] David E. Gerber, Prior cancer does not adversely affect survival in locally advanced lung cancer: A national SEER-medicare analysis, Lung Cancer, May 2016.

[2] Suzuki K., “false-positive reduction in computer-aided diagnostic scheme for detecting nodules in chest radiographs”,academic radiology, volume 13, number 10, pp.10-15, february 2005.

[3] R.H. Chan, Chung-Wa Ho, M. Nikolova “Salt-and-pepper noise removal by median-type noise detectors and detail-preserving regularization”, IEEE Transactions on Image Processing Vol. 14, No. 10, oct 2005.

[4] Priyanka Basak, Asoke Nath, “Detection of Different Stages of Lungs Cancer in CT-Scan Images using Image Processing Techniques”, International Journal of Innovative Research in Computer and Communication Engineering, Vol. 5, No. 5, may2017.

[5] Aniket Gaikwad, Azharuddin Inamdar, Vikas Behera, ” Lung cancer detection using digital Image processing On CT scan Images” International Research Journal of Engineering and Technology, Vol. 03 No. 04 Apr-2016.

[6] Neha Panpaliya, Neha Tadas, Surabhi Bobade, Rewti Aglawe, Akshay Gudadhe “A SURVEY ON EARLY DETECTION AND PREDICTION OF LUNG CANCER”, International Journal of Computer Science and Mobile Computing Vol. 4, No. 1, January 2015.

[7] P. Aggarwal, R. Vig and H.-K. Sardana, Semantic and Content-Based Medical Image Retrieval for Lung Cancer Diagnosis with the Inclusion of Expert Knowledge and Proven Pathology, In proc. of the IEEE second international conference on Image Information Processing ICIIP'2013.

[8] J. Dehmeshki, H. Amin, M.-V. Casique and X. Ye, “Segmentation of pulmonary nodules inthoracic CT scans: A region growing approach”, IEEE Trans. on Med. Imaging, Vol. 27, pp. 467480, 2008.

[9] K. Devaki, V. MuraliBhaskaran and M. Mohan, “Segment Segmentation in Lung CT Images- Preliminary Results”, Special Issue of International Journal on Advanced Computer Theory and Engineering (IJACTE), Vol. 2, No. 1, pp. 84-89, 2013.

[10] Y. Guo, C. Zhou, H.-P. Chan, A. Chughtai, J. Wei, L.-M. Hadjiiski and E.-A. Kazerooni, “Automated iterative neutrosophic lung segmentation for image analysis in thoracic computed tomography”, Med. Phys., Vol.40, No. 8, 2013.

[11] D. Kornack and P. Rakic, “Cell Proliferation without Neurogenesis in Adult Primate Neocortex,” Science, vol. 294, Dec. 2001.

[12] Nunes É.D.O., Pérez M.G., Medical Image Segmentation by Multilevel Thresholding Based on Histogram Difference, presented at 17th InternationalConference on Systems, Signals and Image Processing, 2010.

1

[S1]SAJET office will update with volume, Issue and page numbers