![]() ![]() Moreover, the features extracted from the predicted 3D shapes lead to a higher classification accuracy for six red blood cell types than the features of the 2D image alone. We demonstrate the power of our approach by showing that the 3D shapes of red blood cells can be reconstructed more realistically than with simpler 3D models. Our SHAPR autoencoder is trained with hundreds of 3D shapes and corresponding 2D sections of red blood cells and differentiated induced pluripotent stem cells, and fine tuned with an adversarial network inspired discriminator. We here show that deep learning can be used to predict the 3D shape of single cells and single nuclei from 2D images and thereby reconstruct relevant morphological information. An application to biomedical imaging, where the tradeoff between resolution and throughput is key, is missing so far. Recently neural networks have been proposed to solve the same task and trained to reconstruct the 3D shape of natural objects from 2D photographs. Reconstructing the shapes of three dimensional (3D) objects from two dimensional (2D) images is a task our brain constantly and unnoticeably performs. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |