Intelligent remote medical imaging
Loading...
Date
2020
Journal Title
Journal ISSN
Volume Title
Publisher
UMT.Lahore
Abstract
Many patients die every year due to lack of treatment or error in diagnosis of a certain disease. Developing an artificial intelligence based detection system can help in early diagnosis and treatment of many diseases. In this paper, we present a potential approach to having a collection of disease detection mechanisms using deep learning which utilize medical imaging for classification. To demonstrate this experiment we use two of the most major diseases detectable using medical imaging: Tuberculosis (TB) and Breast Cancer. We have used CNN architecture and compared the erformance of the network using three different activation functions. In the case of Breast Cancer, we used the INbreast dataset which has a total of 115 cases (410 images) of which 90 cases are from women with both breasts (4 images per case) and 25 cases are from mastectomy patients (2
images per case). Our network achieves an accuracy of 94.51% in predicting the cancer in breast. In the case of TB, we used the Shenzhen dataset which has 326 normal x- rays and 336 abnormal x-rays showing various manifestations of tuberculosis. Our network achieves an accuracy of 96% in predicting TB in chest X-rays. To further demonstrate our experiment we have created an API (Application Programming Interface) for both of our detection programs which uses the Flask framework on Python. Using an API we can utilize these detection programs in any OS (operating system), program or device which can be connected to the internet. For public consumption we’ve created a simple front-end web and mobile application based n PHP and Swift languages, respectively.