Fully automated quantitative cephalometry using convolutional neural networks

It is demonstrated that CNNs, which merely input raw image patches, are promising for accurate quantitative cephalometry, and high anatomical type classification accuracy for test set is demonstrated.