Byeon Younghwa

TL;DR: In this paper, a method for analyzing an image of a facial area may comprise the steps of: obtaining an image, dividing the image into a plurality of detailed images using first machine learning algorithm learned based on anatomical inclusion relationship of the facial area; and distinguishing anatomical compositions included in at least one detailed image using second machine learning method based on the anatomical inclusion relationships of the at least 1 detailed image when an input for selecting at least the plurality of the detailed images is received.

TL;DR: A bone modeling apparatus according to an embodiment comprises: an input unit receiving tomographic images with respect to each point of a plurality of adjacent bones; an analysis unit labeling a bone region included in each image of each point with any one of a first region representing a first bone and a second region representing another bone which is medically distinct from the first bone; a correction unit which calculates whether the first region is increased or decreased in relation to the second bone and whether the second region is decreased in proportion to the first one.

TL;DR: A fully automated method using a 2D convolutional neural network to identify superior and inferior vertebrae in a single slice of CT images, and a post-processing for 3D segmentation and separation of cervical vertebraes demonstrated that it achieved comparable accuracies with inter- and intra-observer variabilities of manual segmentation by human experts.