A GMM Based Algorithm To Generate Point-Cloud And Its Application To Neuroimaging

Recent years have witnessed the emergence of 3D medical imaging techniques with the development of 3D sensors and technology. Due to the presence of noise in image acquisition, registration is a must, yet incur error which propagates to the subsequent analysis. An alternative way to analyze medical imaging is by understanding the 3D shapes represented in terms of point-cloud. Though in the medical imaging community, 3D point-cloud is not a “go-to” choice, it is a “natural” way to capture 3D shapes. Another hurdle presented in applying deep learning techniques to medical imaging is the lack of samples. A way to overcome this limitation is by generating samples using GAN like schemes. However, due to different modality in medical images, standard generative models can not be applied directly. In this work, we use the advantage of the 3D point-cloud representation of medical images and propose a Gaussian mixture model based generation and interpolation scheme. For interpolation, given two 3D structures represented as point-clouds, we can generate point-clouds in between, and the experimental validation shows the goodness of the interpolated samples. We also generate new point-clouds for subjects with and without dementia and show that the generated samples are indeed closely matched to the respective training samples from the same class.