ABSTRACT Advancement in technology has changed the viewpoint of the public towards the medical field. Patients believe in the correct diagnosis of a deformity/defect improper functioning before the commencement of actual treatment. Presentation of imaging strategies makes ease in doctors’ life. Now, they can provide satisfactory treatment to patients without performing hit and preliminary trials. The 2-Dimensional (2-D) imaging techniques give a clear idea of simple bone fractures but fail in visualising multiple fractures in a bone. 3-Dimensional (3-D) techniques overcome the challenges of 2-D imaging. In case of any injury such as fracture, arthritis, etc, doctor’s mainly recommended Computerised Tomography or Magnetic Resonance Imaging of the affected part for the correct diagnosis. Computed Tomography scan is a collection of multiple image slices of a target body part. These slices provide a fair idea of intra-articular fractures. Magnetic Resonance Imaging deals with soft tissues for the correct diagnosis of defects. Computed Tomography scan exposes a patient with high ionising radiations of X-Rays, which make a patient more prone to cancer. Moreover, Computerised Tomography and Magnetic Resonance Imaging are unaffordable for the economically weaker section of society due to their high cost. The drawback of 3-Dimaging techniques motivates researchers to develop a model that accept X-Ray image as an input and yields 3-D view at all possible angels from 0° to 360° for a fed image. To serve this purpose, authors propose an innovative Conditional Generative Adversarial network machine learning model that takes input of an X-Ray image and predicts its multi-view.
Paper
Full text
Conversion of two dimensional images into multi-view images of bone using deep learning
Semantic Scholar · Computer Science · 2020
Abstract
ABSTRACT Advancement in technology has changed the viewpoint of the public towards the medical field. Patients believe in the correct diagnosis of a deformity/defect improper functioning before the commencement of actual treatment. Presentation of imaging strategies makes ease in doctors’ life. Now, they can provide satisfactory treatment to patients without performing hit and preliminary trials. The 2-Dimensional (2-D) imaging techniques give a clear idea of simple bone fractures but fail in visualising multiple fractures in a bone. 3-Dimensional (3-D) techniques overcome the challenges of 2-D imaging. In case of any injury such as fracture, arthritis, etc, doctor’s mainly recommended Computerised Tomography or Magnetic Resonance Imaging of the affected part for the correct diagnosis. Computed Tomography scan is a collection of multiple image slices of a target body part. These slices provide a fair idea of intra-articular fractures. Magnetic Resonance Imaging deals with soft tissues for the correct diagnosis of defects. Computed Tomography scan exposes a patient with high ionising radiations of X-Rays, which make a patient more prone to cancer. Moreover, Computerised Tomography and Magnetic Resonance Imaging are unaffordable for the economically weaker section of society due to their high cost. The drawback of 3-Dimaging techniques motivates researchers to develop a model that accept X-Ray image as an input and yields 3-D view at all possible angels from 0° to 360° for a fed image. To serve this purpose, authors propose an innovative Conditional Generative Adversarial network machine learning model that takes input of an X-Ray image and predicts its multi-view.