Detection of Lung Carcinoma using Fuzzy and ACO Techniques

Detection of Lung Carcinoma using Fuzzy and ACO Techniques

Sandeep Kumar Saini M.tech Scholar ECE Department

Gaurav

Assistant Professor Ece Department

Amita Choudhary Assistant professor ECE Department

SUS Tangori, Mohali Punjab SUS Tangori, Mohali Punjab SUS Tangori, Mohali Punjab

Abstract: Lung carcinoma is a disease that occurs because of uncontrolled cell growth in tissues of the lung. Identifying the lung carcinoma involves lot of steps including the segment verification and hot spot analysis. The detection system will focus on identifying the areas of the cancer on the basis of the trained set. In this paper we will focus on identification process by training the data set on the basis of which classification would be done and the same will be then used for testing using fuzzy logic on the basis of the features extracted and ACO.

Keywords: Lung Cancer, Feature Extraction, Fuzzy Logic, ACO

1. INTRODUCTION

   Lung carcinoma is a disease that occurs because of uncontrolled cell growth in tissues of the lung. This growth may lead to metastasis, which is the invasion of adjacent tissue and infiltration beyond the lungs. Treatment and prognosis depend on the histological type of cancer, the stage (degree of spread), and the patient's performance status. Possible treatments include surgery, chemotherapy, and radiotherapy. Survival depends on stage, overall health, and other factors, but on the contrary only 14% of people diagnosed with lung cancer survived 5 years post detection. Symptoms that may suggest lung cancer includes:

   1. chronic coughing or change in regular coughing pattern,

   2. chest pain or pain in the abdomen,

   3. cachexia (weight loss, fatigue, and loss of appetite),

   4. dysphonia (hoarse voice),

   5. clubbing of the fingernails (uncommon)

   6. dysphasia(difficulty swallowing),

   7. Pain in shoulder, chest, arm, wheezing,

   8. Bronchitis or pneumonia,

   9. Decline in Health and unexplained weight loss.

   The earlier presented techniques lack the information distribution capability in terms with focusing on how much the user has been affected and what would be survival rate of the patient in that particular scenario. The sole aim in this is to create a database which would emphasize on various parameters like survival rate of the patient in addition to the existing details main focus being on the system accuracy which will be carried out by using sample images of persons being affected by cancer in comparison to normal people. This paper is divided into sections. Section II discusses about the feature analysis, Section III focuses on Methodology and in section IV results and simulations are discussed.

2. FEATURE ANALYSIS

Image features extraction stage is an important stage that uses algorithms and techniques to detect and isolate various desired portions or shapes (features) of a given image. To predict the probability of lung cancer presence, binarization approach is used. For carcinoma detection feature extraction is a vital step and for this purpose there are two techniques available Binarization approach and Gray level co- occurrence matrix method. In this Gray level co-occurrence matrix has been considered.

1. Binarization Approach

   In this technique, number of white pixels is extracted and is checked against threshold for the normal and abnormal lungs.

   Input output

   Figure 1: Histogram Equalization on DICOM scan image

   If the number of the white pixels of a new image is less than that of the threshold, then it indicates that the image is normal, else, indicates that the image in abnormal.

   2. GREY LEVEL CO-OCCURRENCE MATRIX APPROACH

   Gray-level co-occurrence matrix is created from images using

   graycomatrix function in MATLAB.

   The following features are extracted from various images data base using this method:

   • Contrast

• Energy

• Entropy

• Homogeneity

• Maximum Probability

• Correlation

• Cluster shade

• Cluster Prominence

• Dissimilarity

• Autocorrelation

• Sum variance

• Sum Entropy

• Difference Variance

• Entropy

  III METHODOLOGY

  The first step is the collection of DICOM images (normal and abnormal) from the available database. The next step will be to apply Histogram-Equalization for image enhancement, in order to get the best level of quality. Post enhancing the image, the general features are extracted which gives indication about normal and abnormal image. These features are then compared with the data base. Lastly Fuzzy and ACO techniques are applied for identification of the the stage and to predict the survival rate of the patients.

  Figure 2: Block Diagram of Project methodology

  IV RESULTS AND DISCUSSIONS

  1. Algorithm STEP 1: Upload Image

STEP2: Training using feature extraction

STEP 3: Upload image for classification STEP 4: Extract Properties

STEP 5: Call fuzzy

STEP 6: If accuracy .parameter~=stable STEP 7: Call ACO

STEP 8: Publish result

Figure 3: representing uploaded base image

The above figure represents the uploaded image. The figure has all the entire data which is uploaded i.e. the classified and not classified data.

Figure 4: extracted region in the lung cancer image.

Figure 5: 2nd Stage cancer image with message box

Figure 6: 3rd Stage cancer image with message box

Figure 5 and figure 6 show the 2nd and 3rd stage cancer. In this, image feature extraction is done to detect and isolate various desired portions or shapes (features) of the image. Both the images depict the abnormal lungs. In accordance with the output, Ant colony optimization techniques are further called to detect carcinoma stages.

Table 1: Extracted features of Dicom images

Table2:Classification of Survival Rate

Table 3: Summarize Result

V CONCLUSION

This paper concludes the lung carcinoma detection mechanism. There are two stages of detection. One is called training and another is called testing phase. In the training phase, Feature extraction is done with the help of the gray level co-occurrence method. In the classification stage a combination of two classification methods namely FUZZY and ACO has been used. The result shows the optimum and significant growth. The current research work can be further carried out on different algorithms such as BFO and it may be applied to different types of cancer.

REFERENCES

1. Chuen Chien Lee, Fuzzy logic in control systems i.e. fuzzy logic controller, IEEE Transactions on Systems, Man and cybernetics, Vol 20, No.2, March/April 1990.

2. Reza Langari, Past, present and future of fuzzy control: A case for aplication of fuzzy logic in hierarchical control, IEEE, pp.760-765, 1999.

3. Michael Phillips, Nasser Altorkic, John H.M. Austind, Peter Schmittk ,Prediction of lung cancer using volatile biomarkers in breath ,Proceedings of USA Cancer Biomarkers, pp: 95 109,2007.

4. M. A. Saleem Durai, S. N. Iyengar and A. Kannan ,Enhanced Fuzzy Rule Based Diagnostic Model for Lung Cancer using Priority Values, International Journal of Computer Science and Information Technologies,pp:707-710,Vol-2,2011.

5. Fatma Ayari, Mekki Ksouri and Ali Alouani, A computer based model for lung cancer analysis, IJCSI International Journal of Computer Science Issues, Vol. 9, Issue 5, September 2012.

6. Ada and Rajneet Kaur, Feature Extraction and Principal Component Analysis for Lung Cancer Detection in CT scan Images, International Journal of Advanced Research in Computer Science and Software Engineering, ISSN: 2277- 2281, Vol-3, Issue 3, March 2013.

7. Gurpreet Kaur and Jaspreet Kaur, Fuzzy logic based multi thresholding for Lung cancer detection, pp: 978-983, Sept 2013.

8. Ada and Rajneet Kaur, Early Detection and Prediction of Lung Cancer Survival using Neural Network Classifier, International Journal of Application or Innovation in Engineering & Management, pp: 2319 4847, Vol-2, Issue 6, June 2013.

9. Ada and Rajneet Kaur Using Some Data Mining Techniques to Predict the Survival Year of Lung Cancer Patient International Journal of Computer Science and Mobile Computing IJCSMC, pp: 1-6, Vol- 2, Issue. 4, April 2013.

10. K.A.G. Udeshani, R.G.N. Meegama, T.G.I. Fernando, Statistical Feature-based Neural Network Approach for the Detection of Lung Cancer in Chest X-Ray Images, International Journal of Image Processing (IJIP), Volume-5, Issue -4 , 2011.

11. Mokhled S. AL-TARAWNEH, Lung Cancer Detection Using Image Processing Techniques, Leonardo Electronic Journal of Practices and Technologies, Issue 20, January-June 2012.

12. Taranpreet Singh Ruprah, Face Recognition Based on PCA Algorithm, Special Issue of International Journal of Computer Science & Informatics (IJCSI), 22315292, Vol.-II, Issue-1, 2.

Mr. Sandeep Kumar Saini is pursuing M.Tech. in Electronics and Communication Engineering from SUS Tangori. His current field of working is in Detection of Lung carcinoma using fuzzy and ACO Techniques.