Deep Learning Model: Can we Ignore Foreign Objects in Chest X-ray Screening?
Lung (Pulmonary) abnormalities, such as Tuberculosis (TB), Asthma and/or Chronic obstructive have been global threats for years. According to the World Health Organization's (WHO) report of 2019, nearly 1.5 million people have died from Tuberculosis alone. Computer scientists have worked together with medical experts to design automated screening systems for chest X-ray (CXR) images. However, in most of the research work, detection of foreign objects, such as buttons, coins, ring, pins, bone pieces and also other medical devices (like a pacemaker) hasn't been considered, which have hindered the performances of automated screening system. The circular foreign objects, such as coins can often be confused with nodules, which is one of the primary indicators of Tuberculosis. Thus, for an automated screening process, we need to separate such foreign objects. This research is mainly focused on the detection of foreign objects that are of almost all shapes, sizes and texture in CXRs using a convolutional neural network. So, unlike the prior works, we will be using deep learning models, such as Faster R-CNN (Faster Region Proposal Convolutional Neural Network), to detect the foreign objects in the CXRs. That is how, instead of relying on handcrafted features, we now let the machine to extract the automated and distinguished features, to achieve minimum error possible (technically, 10^-4). Also, we localize their spatial position in CXR, so that the further process of screening can be advanced and at the same time misdiagnosis and confusion can be eliminated.
Neupane, Amul, "Deep Learning Model: Can we Ignore Foreign Objects in Chest X-ray Screening?" (2020). IdeaFest. 158.