Skip to main content
. Author manuscript; available in PMC: 2020 Jun 16.
Published in final edited form as: Int J Comput Assist Radiol Surg. 2019 Apr 20;14(9):1463–1473. doi: 10.1007/s11548-019-01975-5

Fig. 1.

Fig. 1

Schematic representation of the convolutional neural network used in this work. A single input image is processed by multiple stages of convolutional and pooling layers, resulting in a stack of belief maps, where each map corresponds to a landmark location. During the stage-wise application, these belief maps are refined