3. INTRODUCTION
• Tuberculosis is an infectious diseases caused by the bacillus Mycobacterium
tuberculosis, which typically affects the lungs
• TB is the 2nd leading cause of death from an infectious disease worldwide
• Accurate diagnosis is the key to controlling the disease
• We present an automated approach to detect tuberculosis using conventional
posteroanterior chest radiographs
• Typical manifestations of TB in chest X-rays are infiltrations, cavitations,
effusions or miliary patterns
• Focus is on segmentation techniques, extraction of features and classification as
normal or abnormal on the basis of classifier i.e. trained on number of features
4. LITERATURE SURVEY
Poornimadevi.CS, Helen Sulochana.C
• In the existing method, cavity detection, ribs and diaphragm elimination is
difficult to examine TB in chest radiographs.
• To overcome the difficulties lung region is extracted by using registration based
segmentation method.
• Segmentation of lung region is performed after the registration process to handle
complex segmentation problems.
• In this paper, they describe how to discriminate between the normal and abnormal
CXRs using image processing techniques.
• Contrast enhancement and noise removal can be performed by using the
preprocessing techniques.
5. Laurens Hogeweg, Rodney Dawson, Grant Theron
• Computer aided detection system was developed which combines the results of
textural, shape and focal analysis.
• The lung and the clavicle segmented to limit the analysis by the subsystems to the
lung fields and provide them spatial context.
• Virtual collimation procedure is used which yields images with standardised lung
images, improving robustness of subsequent analysis.
• Abnormality detection system based on the textural, focal and shape.
• All subsystems are introduced in one for easy detection.
6. Anju Mathew, Athira V.R
• This paper proposes an efficient diagnosis of tuberculosis with the aid of chest
radiographs.
• At first they use content based image retrieval technique through we can identify
the lung CXR images by partial radon transform.
• They compute a lung model based on deformable non registration algorithm
• Extract the lung region from the chest radiographs using SIFT flow algorithm
followed by graph cut lung segmentation method.
• Then extract various features and by using SVM classifier the x-rays are classified
as TB affected or not.
• It also provides information whether the patient is highly TB affected or TB
starting level or not TB.
7. Stefan Jaeger, Alexandros Karargyris, Sema Candemir, Les Folio, Jenifer
Siegelaman, Fiona Callaghan
• This system first segments the lung region using a graph cut optimisation method
• This method combines intensity information with personalised lung atlas models
derived from a training set
• For the segmented lung field, the system then computes a set of shape, edge and
texture features as input to a pre-trained binary classifier
• Finally, using decision rules and thresholds, the classifier outputs whether the
input CXR is TB positive or not
8. Stefan Jaeger, Alexandros Karargyris, Sameer Antani, George Thoma
• Automated approach for TB detection on conventional posteroanterior chest
radiographs
• Segmentation mask for lung is achieved by a combination of lung shape model,
segmentation mask, and a simple intensity model
• JSRT data for training the lung model
• Binary alignment of the lung shaped model to map the model into an input x-ray
11. SOFTWARE PACKAGE
• MATLAB is a high performance language for technical computing
• It allows matrix manipulations, plotting of functions and data, implementation of
algorithms, creation of user interfaces and interfacing with programs written in
other languages, including C, C++, C#, JAVA, Fortran and Python
• It integrates computation, visualization and programming in an easy-to-use
environment where problems and solutions are expressed in familiar mathematical
notations
14. mask = zeros(size(picture));
mask(roi(2):(roi(2)+roi(4)),roi(1):(roi(1)+roi(3))) =0;
bw = activecontour(ims,mask,'edge');
figure;
imshow(bw)
imcc=imsubtract(B_close,imb);figure;
imshow(imcc)
bw1 = bwlabel(B_close, 8);
stats = regionprops(bw1, 'BoundingBox', 'Centroid');
hold on
for object = 1:length(stats)
bb = stats(object).BoundingBox;
bc = stats(object).Centroid;
15. rectangle('Position',bb,'EdgeColor','r','LineWidth',2)
plot(bc(1),bc(2), '-m+')
a = text(bc(1)+15,bc(2), strcat('X: ', num2str(round(bc(1))), ' Y: ', num2str(round(bc(2)))));
set(a, 'FontName', 'Arial', 'FontWeight', 'bold', 'FontSize', 12, 'Color', 'yellow');
end
hold off
bw2 = bwlabel(imcc);
stats = regionprops(bw2,'BoundingBox', 'centroid');
bb = stats(object).BoundingBox;
bc = stats(object).Centroid;
rectangle('Position',bb,'EdgeColor','r','LineWidth',2)
plot(bc(1),bc(2), '-m+')
a = text(bc(1)+15,bc(2), strcat('X: ', num2str(round(bc(1))), ' Y: ', num2str(round(bc(2)))));
set(a, 'FontName', 'Arial', 'FontWeight', 'bold', 'FontSize', 12, 'Color', 'yellow');
end
hold off
16. IMPLEMENTATION
• A grayscale image is a digital image, in which the value of each pixel is an
individual sample.
• Grayscale images are the result of measuring the intensity of light at each pixel in
a single band of the light spectrum.
17. • Region of interest is a portion of an image that we want to filter or perform some
other operation.
• Here the ROI is defined by using the rectangular box.
18. • The portion of image from the region of interest is subtracted from the gray scale
image.
• To convert the image into binary image thresholding is done for the subtracted
image.
19. • Closing is used to fill gaps in an image. imclose() performs morphological closing
on the gray scale or binary image.
• The morphological close operation is a dilation followed by an erosion using the
same structuring element for both the operations.
• imopen() performs morphological opening on the gray scale on the binary image.
Morphological open operation.
20. • regionprops() returns measurements for the set of properties specified by
properties for each 8-connected component (object) in the binary image.
• Centroid indicates the center of mass of the region.
• Bounding box is the smallest rectangle containing the region.
21. RESULT AND DISCUSSION
• TB has diverse manifestations and to analyse CXRs automatically, algorithms that
focus on different manifestations need to be combined
• For a given input chest X-ray, we first segment the lung field and then extract a set
of features for shapes, curvatures and textures from the segmented lung field
• Using the extracted features we classify whether the input CXR is normal or not
22. FUTURE SCOPE
• A system that can assist radiologist and public health providers in the screening
and decision process
23. ADVANTAGES AND DISADVANTAGES
Advantages
• Devoid of inter-reader variability
• Standardised way of reporting
• Objective and reproducible results
• Fast and effective
Disadvantages
• Poor specificity
24. CONCLUSION
• We are developing an automated system that screens chest X-rays for
manifestations of TB .
• When the input is given as the chest x-ray image it thereby detect the presence of
cavity.
• It indicates the presence or absence of tuberculosis
25. WORK SCHEDULE
Project start 18 July 2017
Topic selection 27 July 2017
Literature survey 2 August 2017
Zeroth review 18 August 2017
Data collection 27 September 2017
First review 4 October 2017
Implementation 1 December 2017
Report preparation 2 January 2018
Final review 4 April 2018
26. REFERENCES
[1] Anju Mathews, Athira V.R., “Detection of Tuberculosis Using Chest X-Rays”,
International Journal of Advanced Research in Electronics and Communication
Engineering, Volume 4, Issue 6, June 2015.
[2] Hrudya Das ,Ajay Nath ,"An Efficient Detection of Tuberculosis from Chest X-
rays", International Journal of Advance Research in Computer Science and
Management Studies, Volume 3, Issue 5,May 2015.
[3] Poornimadevi C.S., Helen Sulochana C., “Automatic Detection of Pulmonary
Tuberculosis using Image Processing Techniques”, IEEE WiSPNET 2016
Conference.
27. [4] S. Jaeger, A. Karargyris, S. Candemir, J. Siegelman, L. Folio, S. Antani and G.
Thoma, “Automatic tuberculosis screening using chest radiographs,’’
Quantitative Imaging in Medicine and Surgery, vol. 3, 2013.
[5] A. Karargyris, S. Antani, and G. Thoma, “Segmenting anatomy in chest x-rays for
tuberculosis screening,” in 33rd Int. Conf. of the IEEE Engineering in Medicine
and Biology Society (EMBS), 2011.
[6] World Health Organization. Global tuberculosis control: WHO report 2010. World
Health Organization, 2010.
