arXiv reaDer
Bodypart Regressorを介した登録のためのCTスキャンの効率的な事前調整
Efficient Prealignment of CT Scans for Registration through a Bodypart Regressor
Convolutional neural networks have not only been applied for classification of voxels, objects, or images, for instance, but have also been proposed as a bodypart regressor. We pick up this underexplored idea and evaluate its value for registration: A CNN is trained to output the relative height within the human body in axial CT scans, and the resulting scores are used for quick alignment between different timepoints. Preliminary results confirm that this allows both fast and robust prealignment compared with iterative approaches.
updated: Thu Sep 19 2019 10:02:58 GMT+0000 (UTC)
published: Thu Sep 19 2019 10:02:58 GMT+0000 (UTC)
参考文献 (このサイトで利用可能なもの) / References (only if available on this site)
被参照文献 (このサイトで利用可能なものを新しい順に) / Citations (only if available on this site, in order of most recent)アソシエイト