International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 04 Issue: 04 | Apr -2017 www.irjet.net p-ISSN: 2395-0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 3407 Classification of benign and malignant lung nodules using image processing techniques Moffy Crispin Vas 1 , Amita Dessai 2 1 Student, Dept. of ETC, Goa Engineering College, Goa, India 2 Assistant Professor, Dept. of ETC, Goa Engineering College, Goa, India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract – Cancer is the second leading cause of most number of deaths worldwide after the heart disease, out of which, lung cancer is the leading cause of deaths among all the cancer types. Hence, the lung cancer issue is of global concern and thus this work deals with detection of malignant lung cancer nodules and tries to distinguish it from the benign nodules by processing the Computer tomography (CT) images with the help of Haar wavelet decomposition, Haralick feature extraction followed by artificial neural networks (ANN) Key Words: Computer tomography, Lung cancer, malignant, Haralick features, ANN 1. INTRODUCTION Lung cancer contributes to about 19% of the deaths globally. A person suffering from lung cancer has an overall 5 years of survival with only 15% assurance in developed countries and 5% in developing countries. If the cancerous nodules are detected at an early stage the survival rate can shoot up to 50-60%. Computer tomography scans have been proved useful in detecting lung cancer and hence CT scans have reduced cancer mortality rates by 20% but at the cost of false positive rate of 96%. A lung nodule is the white spot that appears on the CT image having a size of about 3 cms. It becomes difficult to conclude visually whether the nodule is malignant or benign. Around 40% of the lung nodules are cancerous and have to be detected as early as possible to degrade the mortality rates. It becomes necessary to develop an automated system that will classify the malignant nodules at an early stage with increased accuracy and speed. 2. RELATED WORKS Khin Mya Mya Tun extracted geometrical features from the CT images and classified the images using feed forward artificial neural networks [1]. S.A.Patil and used x-ray images and preprocessed them with median filters. Segmentation techniques like region growing and morphological operations were used and they also utilized geometrical and first order statistical texture features for classifying the cancer using ANNs [2]. Anita Chaudhary and used three different image enhancement techniques, out of which they reported that Gabor filter gave them best results. Thresholding, watershed segmentation techniques and extracted features such as area, roundness and eccentricity of lung nodules were used for classifying the lung cancer and its stages [3]. Md. Badrul Alam Miah segmented out the left and right lung separately using edge maps. By extracting 33 different features they classified the images using feed forward neural network [4]. Muhammed Anshad gives a comparative survey of all the methods used for automated cancer detection systems. Comparisons are made based on accuracy, advantages and disadvantages of the methods [5]. Amjed S.AlFahoum designed an automated intelligent system for nodule detection and classification of lung cancer in CT images. In this work they made use of morphological operations and utilized geometrical features for classification [6]. Gawade Prathamesh Pratap carried out p- tile thresholding and watershed processing on PET/CT images followed by the use M type morphology to display cancer image if any with the help of MATLAB [7]. Mohsen Keshani used an active contour for lung segmentation and detected ROIs by stochastic 2D features. They further used