what algorithm will be better in order to identify the tuberculosis in x ray image ??

actually the normal segmentation process is done now i want to use the algorithm and let me know what will be the best .

Answer 1

[Star Strider]

What sort of pulmonary tuberculosis lesion do you want to identify? Miliary tuberculosis? Ghon’s Complex? Rasmussen’s aneurysm?

You need to be more specific. Please consult a Board-Certified Radiologist at your institution to find out how to correctly diagnose M. tuberculosis pulmonary infection on the basis of whatever imaging modality you choose to use (PA and L lateral plain films, CT scan, or something else).

[rahul]

sir actually i am trying to find the tuberculosis from lungs still no specified type i had choose. I had done the gui part and the segmentation part now i am getting struck with the algorithm what i have to use along with the coding part to do the further steps in my projects. suggest some algorithm also sir. If so any idea reg: detecting any particular identify path

[Image Analyst]

Looks like you totally ignored everything Star and I said. Did you read the answers? Have you narrowed down the diagnosis to certain types of lesions? Are you partnering with a radiologist yet? Did you get more specific in your problem definition in any way at all? Did you search VisionBib (or other sources) to find published algorithms that you can program up, like I suggested? It looks like all you did was to paraphrase your original question, and not do anything we suggested yet. At this point, your question is not even a MATLAB programming question since there's no code that you've uploaded. In this forum we help with MATLAB code (errors, syntax, simple program flow), not complicated algorithm development.

[Star Strider]

I completely understand the problem you are facing. Diseases are pleomorphic (the same disease does not always appear the same way in different individuals, for a variety of reasons), and other diseases (in this instance, other than tuberculosis) can present similarly. For instance, miliary tuberculosis can appear quite similar to berylliosis or sarcoidosis. Even with advanced image processing software to assist them (since most radiographs are digital these days), it is frequently difficult for humans to make accurate diagnoses on the basis of a radiograph alone. As I mentioned, tuberculosis itself can have many different appearances.

That is the reason I suggest that you consult with an appropriately-educated radiologist to discuss what the radiologist looks for to suggest to the the treating clinician the possible diagnosis of pulmonary tuberculosis. That will guide your algorithm development.

If you haven’t seen this article, you might find it helpful: ‘Two public chest X-ray datasets for computer-aided screening of pulmonary diseases’, Quant Imaging Med Surg. 2014 Dec;4(6):475-7. It’s a recent, full-text, free article. The emphasis is on the diagnosis of pulmonary tuberculosis.

This one is not free, but you likely have access to it through your university library: ‘Automatic tuberculosis screening using chest radiographs.’ IEEE Trans Med Imaging. 2014 Feb;33(2):233-45

I always suggest a PubMed search, which is where I found the articles I cited here.

I completely agree with Image Analyst’s comment.

Answer 2

[Image Analyst]

Perhaps look over section 20.7.2 here: http://www.visionbib.com/bibliography/contentsmedical.html#Medical%20Applications,%20CAT,%20MRI,%20Ultrasound,%20Heart%20Models,%20Brain%20Models