CCS-GAN: COVID-19 CT Scan Generation and Classification with Very Few Positive Training Images.

We present a novel algorithm that is able to generate deep synthetic COVID-19 pneumonia CT scan slices using a very small sample of positive training images in tandem with a larger number of normal images. This generative algorithm produces images of sufficient accuracy to enable a DNN classifier to achieve high classification accuracy using as few as 10 positive training slices (from 10 positive cases), which to the best of our knowledge is one order of magnitude fewer than the next closest published work at the time of writing. Deep learning with extremely small positive training volumes is a very difficult problem and has been an important topic during the COVID-19 pandemic, because for quite some time it was difficult to obtain large volumes of COVID-19-positive images for training. Algorithms that can learn to screen for diseases using few examples are an important area of research. Furthermore, algorithms to produce deep synthetic images with smaller data volumes have the added benefit of reducing the barriers of data sharing between healthcare institutions. We present the cycle-consistent segmentation-generative adversarial network (CCS-GAN). CCS-GAN combines style transfer with pulmonary segmentation and relevant transfer learning from negative images in order to create a larger volume of synthetic positive images for the purposes of improving diagnostic classification performance. The performance of a VGG-19 classifier plus CCS-GAN was trained using a small sample of positive image slices ranging from at most 50 down to as few as 10 COVID-19-positive CT scan images. CCS-GAN achieves high accuracy with few positive images and thereby greatly reduces the barrier of acquiring large training volumes in order to train a diagnostic classifier for COVID-19.

Computational observers and visualization methods for stereoscopic medical imaging.

As stereoscopic display devices become common, their image quality assessment evaluation becomes increasingly important. Most studies conducted on 3D displays are based on psychophysics experiments with humans rating their experience based on detection tasks. The physical measurements do not map to effects on signal detection performance. Additionally, human observer study results are often subjective and difficult to generalize. We designed a computational stereoscopic observer approach inspired by the mechanisms of stereopsis in human vision for task-based image assessment that makes binary decisions based on a set of image pairs. The stereo-observer is constrained to a left and a right image generated using a visualization operator to render voxel datasets. We analyze white noise and lumpy backgrounds using volume rendering techniques. Our simulation framework generalizes many different types of model observers including existing 2D and 3D observers as well as providing flexibility to formulate a stereo model observer approach following the principles of stereoscopic viewing. This methodology has the potential to replace human observer studies when exploring issues with stereo display devices to be used in medical imaging. We show results quantifying the changes in performance when varying stereo angle as measured by an ideal linear stereoscopic observer. Our findings indicate that there is an increase in performance of about 13-18% for white noise and 20-46% for lumpy backgrounds, where the stereo angle is varied from 0 to 30. The applicability of this observer extends to stereoscopic displays used for in the areas of medical and entertainment imaging applications.