3 ms·
3D reconstruction using single X-ray image for Pediatric Orthopedic applications
- coolwulf 7y agoAuthor here. The motivation behind this project is the need for 3D representation by doctors in Orthopedic assist diagnosis. However the limit on the amount of dose could be used for pediatric patients inhibit the amount of CT scans could be done. There are other systems on the market such EOS Imaging Flex Dose which is based on an atlas based registration method to reconstruct two 2D x-ray projections into 3D spinal representation. However such system is not only very expensive (selling in USA at $1M for one system) but also produce certain image artifacts due to the atlas is based on normal people. This work is to show the feasibility of deep learning could surpass performance of these other systems and achieve high resolution spinal 3D representations using 2D images. Because X-ray is attenuation based. Or we can say image intensity of each pixel is I = I_0*exp(-ut). I_0 is the incoming X-ray fluence, u is the attenuation coefficient, t is the path length for the x-ray. So different contrast and also the contour of the features on the resulting image I bears information of features behind the front vertebrate. These info can be extracted using feature extraction of the deep learning network. With enough data training, these could be learned and this is the basis of this work. In orthopedic, the spinal angle etc. info is more important than the detailed bone artifacts such as small fractures. By combining both AP and lateral x-rays, together with this generated 3D info, it should be sufficient for this kind of application.