Virtual Tuberculosis Detector

Project Title

Virtual Tuberculosis Detector

Project Area of Specialization Artificial IntelligenceProject Summary

In the present work, results of automatic classification of medical images are presented in two categories: with and without tuberculosis.

To our knowledge this study reports first-time results of an active TB case finding strategy based on on-spot symptom screening using a web based application

To carry out the classification, features are extracted using deep learning and the RESNET50 neural network. Cross-validation and the formation of training and test sets were the two classification scenarios used. The scenario with the best results was the one in which the training and test sets were formed with an accuracy greater than 85%.

The classification method that shows the best performance in the two scenarios implemented in this work is SVM. As can be seen in the results obtained in the present work, these far exceed chance and allow to carry out the classification of images in an efficient way.

Computer tomography (CT) of the abdomen, CT of the head, magnetic resonance imaging (MRI) of the brain, and MRI of the spine were all used in this investigation. Our suggested CNN architecture could automatically categorize these 4 sets of medical photos by image modality and anatomic location after converting them to JPEG (Joint Photography Experts Group) format. In both the validation and test sets, we achieved outstanding overall classification accuracy (>99.5 percent)

A test harness is required while constructing a framework for a predictive modeling issue.

The test harness specifies how the domain’s sample of data will be used to assess and compare potential models for a predictive modeling challenge.

The proposed method was developed. The characteristics of the images that are used as classification attributes are extracted with KERAS. KERAS is an open-source neural networks library created in Python that contains the ResNet50 architecture; this architecture will help to extract the characteristics of the images through arrays.

There are several methods to organize a test harness, and there is no one-size-fits-all solution for all applications.

Using a piece of the data for training and tuning the models and a portion for giving an objective assessment of the tuned model’s skill on out-of-sample data is a popular strategy.

A training dataset and a test dataset are created from the data sample. The model is assessed using a resampling approach such as a k-fold pass on the training dataset, and the set may be further separated into test data for tuning the model’s hyperparameters.

The collected results allow us to assess the viability of the methods adopted. It also allows us to identify the best classification scenario and machine learning method to carry out the classification of radiographs with and without tuberculosis.

The data availability underlying the results presented in the study are included within the manuscript.  It was performed as a part of the Employment of Institutions.

Project Objectives

our objectives or goals of this project is or the reason of the app are to assist increment tuberculosis case location and enrollment in treatment through progressed screening, documentation, and referral hones within the private division. It has the potential for quick scale and tall effect within the battle against tuberculosis.

In Pakistan, the standard paper processes for screening, sputum sample collection and testing, and referral between clinical facilities are cumbersome and can result in delays in decision making, follow-up, and care. To address these challenges, our FYP project has developed an application which to replace paper forms with a more efficient digital platform and support the reporting needs of clinical providers such as doctors and nurses, laboratory scientists, pharmacists, and patent and proprietary drug vendors. 

WHO gauges that approximately one-third of all occurrence cases of dynamic TB are not properly analyzed or get care of the flawed quality exterior of national TB programs and are not being notified. Among those cases who are in the long run analyzed, the delay is regularly long. There's abundant direct prove from national overviews of TB prevalence and other research that the pool of undetected TB cases remains expansive large expansive huge in many countries in spite of the scaling up and decentralizing of TB conclusion and treatment, especially in certain chance bunches, such as individuals living with HIV, 31 close contacts of individuals with TB, miners, prisoners, homeless people, and a few clinical chance bunches so that's why Our targets are to identify the number of TB-related apps accessible within the fundamental app stores, portray their characteristics, assess they extend of functionalities, recognize any thorough testing of the accessible apps, and filter.

basically, our objectives or goals are:

diminishing the chance of destitute treatment results, wellbeing sequelae, and the adverse social and financial results of TB for the person. This reduces enduring, the predominance of TB, and passing from TB.
 lessening TB transmission by shortening the term of infectiousness. This decreases the rate of TB contamination and thus contributes to the diminished frequency of TB infection.
A second objective is to run the show out dynamic malady to assist recognize individuals who are qualified for the treatment of idle TB contamination for illustration, among people living with HIV and contacts who are more youthful than 5 years.

Combining screening for TB with screening for TB chance variables can moreover help map person or community-level hazard variables and socioeconomic determinants that have to be tended to viably anticipate the disease.

Our Mission to Viable conclusion the TB scourge in Pakistan that's why  Our assessment appears that more refined work is required in this zone. Including TB patients in treatment within the plan of these apps is prescribed.

Project Implementation Method

We created a demonstration of TB transmission, care-seeking behavior, and diagnostic/treatment hones in Pakistan and investigated the effect of six diverse rollout procedures. Giving Xpert to 40% of public-sector patients with HIV or earlier TB treatment (comparative to current national technique) decreased TB frequency by 0.2% (95% instability run [UR]: ?1.4%, 1.7%) and MDR-TB frequency by 2.4% (95% UR: ?5.2%, 9.1%) relative to existing hone but required 2,500 extra MDR-TB medications and 60 four-module GeneXpert frameworks at most extreme capacity. Assist counting 20% of unselected symptomatic people within the open division required 700 frameworks  >2,200 frameworks and dependable treatment referral. These discoveries are subject to significant instability concerning private-sector treatment designs, understanding care-seeking behavior, indications, and infectiousness over time; these vulnerabilities ought to be tended to by future investigation.

Model Structure: TB Diagnosis, Treatment, and Care-Seeking
MDR-TB, TB Treatment, and HIV
Since Xpert can detect resistance to rifampin, we modeled infection with drug-susceptible and rifampin-resistant TB (as a proxy for MDR-TB) as separate strains, with rifampin resistance propagating through both inadequate treatment (potentially different across sectors) and transmission of resistant strains. 

Model Calibration
We initiated our model at a steady-state, reflecting trends in India before 2005. Taking all other parameters as fixed, we fitted the following model parameters the transmission rate (number of transmitted infections per person-year), calibrated to the WHO-estimated overall TB incidence the duration of infectiousness before symptom onset, calibrated to the WHO-estimated duration of TB disease.

Where available, we assumed Xpert improved diagnosis and treatment as follows (numbers in parentheses denote sensitivity analysis ranges): increased sensitivity for less infectious (smear-negative) TB from 0%–20% (depending on healthcare sector) to 73% (60%–80%); increased sensitivity to highly infectious TB 

Sensitivity Analysis
Some data suggest that TB drug prescriptions are suboptimal and treatment outcomes in the private sector may be inferior compared with the public sector, Therefore, we conducted an alternative analysis where we assumed lower-quality treatment in the private sector. 

Uncertainty Analysis
in addition to the sensitivity analyses above, to obtain measures of uncertainty that also included the impact of the random nature of both the disease and healthcare behavior processes, we built a stochastic version of the model.

Benefits of the Project

Screening for tuberculosis (TB) disease aims to improve early TB case detection. The ultimate goal is to improve outcomes for people with TB and to reduce Mycobacterium tuberculosis transmission in the community through improved case detection, reduction in diagnostic delays, and early treatment. Before screening programs are recommended, evidence is needed of individual and/or community-level benefits.

A major advantage of Xpert MTB/RIF is that it can identify people with drug-resistant TB very rapidly" So it looks as if most patients who needed treatment got it in the end, regardless of how they were diagnosed, and when they followed up the patients six months later to see how well they were doing, they could not detect any difference in death rates, or in their state of health. (They measured morbidity using the Karnofsky performance status index, which scores patients from 100 percent = in perfect health to 0 percent = dead). 

"A major advantage of Xpert MTB/RIF is that it can identify people with drug-resistant TB very rapidly. In this study the proportion of people with drug-resistant TB was low, and there might be a much greater benefit from using Xpert in settings where drug-resistant TB is more common, providing that people identified as having drug-resistant TB start on the correct treatment promptly."

community-level benefits of orderly screening are remarkably restricted, given the tall open wellbeing significance, long history, and scale on which this approach has been actualized within the past. Huge cluster randomized trials, such as the ZAMSTAR study, with long-term follow-up, would be required to supply more proof for such a benefit on the off chance that in fact, it exists, in a perfect world counting thinks about that assess a run of interventions with distinctive screening force in different epidemiological settings

Accurate across patient populations Immunosuppressed. BCG-vaccinated.
Consistent results6,7 Sensitivity: 98.8%/li> Specificity: >99.1%/li>
High test accuracy around result cut-off due to regulatory approved borderline zone – helping to prevent inappropriate therapy
Individual and community-level benefits from active screening for TB disease remain uncertain. So far, the benefits of earlier diagnosis on patient outcomes and transmission have not been established.

Microscopy of sputum smears is simple and inexpensive and allows rapid detection of the most infectious cases of pulmonary TB. Sputum specimens from patients with pulmonary TB, especially those with the cavitary disease, often contain sufficiently large numbers of AFB to be detected by microscopy.

Major advantages of the Xpert MTB/RIF assay are that • Results are available quickly, and

• Minimal technical training is required to run the test. Additionally, the Xpert MTB/RIF assay can quickly identify possible multidrug-resistant TB (MDR TB)

The above are some benefits of the tuberculosis web application.

Technical Details of Final Deliverable

the technical details of our project id that it is based on artificial intelligence. in this, we use machine learning as we know that Machine learning is a branch of computing that studies the design of algorithms with the ability to “learn.” A subfield would be deep learning, which is a series of techniques that make use of deep artificial neural networks, that is, with more than one hidden layer, to computationally imitate the structure and functioning of the human organ and related diseases. The analysis of health interest images with deep learning is not limited to clinical diagnostic use. It can also, for example, facilitate surveillance of disease-carrying objects. There are other examples of recent efforts to use deep learning as a tool for diagnostic use. Chest X-rays are one approach to identify tuberculosis; by analyzing the X-ray, you can spot any abnormalities. A method for detecting the presence of tuberculosis in medical X-ray imaging is provided in this paper. Three different classification methods were used to evaluate the method: support vector machines, logistic regression, and nearest neighbors. Cross-validation and the formation of training and test sets were the two classification scenarios used. The acquired results allow us to assess the method’s practicality. next, we use for the development of our web-based application w use HTML, CSS JAVASCRIPT for frontend development and for backend we used PYTHON and DATABASE OF FIREBASE, storage which we used is FIREBASE STORAGE. for authentication like sign in or sign out we used FIREBASE ANTHENTICATION. and we trained our model on CNN a convolutional neural network (CNN/ConvNet) deep learning algorithm. and we deployed it on the flask. for data saving, we used FIREBASE REAL-TIME DATABASE. for the training of the model we used python language. that above all are the technical details of our web-based application.

For the present work, the firebase database was used, the X-ray images of this database were collected from the tuberculosis control program adopted by set contains 138 radiographs, of which 80 radiographs correspond to healthy patients (normal) and 58 radiographs show manifestations of tuberculosis (abnormal). This database is available. All images have been de-identified and are in DICOM format. The set includes a wide variety of abnormalities, such as spillage patterns. The dataset contains radiological readings in the form of a text file. Each image contains a label that aids with image identification. 

The test harness specifies how the domain’s sample of data will be used to assess and compare potential models for a predictive modeling challenge.

There are several methods to organize a test harness, and there is no one-size-fits-all solution for all applications.

The references or study was carried out in accordance with the recommendations of the ethics manners of the Medical Sciences.

Final Deliverable of the Project Software SystemCore Industry MedicalOther Industries Health Core Technology Artificial Intelligence(AI)Other Technologies Artificial Intelligence(AI), OthersSustainable Development Goals Good Health and Well-Being for People, Life on LandRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	56357
Documentation	Miscellaneous	3	1299	3897
Report	Miscellaneous	2	580	1160
stationery	Miscellaneous	5	260	1300
API	Equipment	1	15000	15000
system	Equipment	1	35000	35000