Abstract | ||
---|---|---|
We present a self-supervised approach to training convolutional neural networks for dense depth estimation from monocular endoscopy data without
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a priori</italic>
modeling of anatomy or shading. Our method only requires monocular endoscopic videos and a multi-view stereo method, e.g., structure from motion, to supervise learning in a sparse manner. Consequently, our method requires neither manual labeling nor patient computed tomography (CT) scan in the training and application phases. In a cross-patient experiment using CT scans as groundtruth, the proposed method achieved submillimeter mean residual error. In a comparison study to recent self-supervised depth estimation methods designed for natural video on
<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in vivo</italic>
sinus endoscopy data, we demonstrate that the proposed approach outperforms the previous methods by a large margin. The source code for this work is publicly available online at
<uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/lppllppl920/EndoscopyDepthEstimation-Pytorch</uri>
. |
Year | DOI | Venue |
---|---|---|
2020 | 10.1109/TMI.2019.2950936 | IEEE Transactions on Medical Imaging |
Keywords | DocType | Volume |
Estimation,Endoscopes,Cameras,Videos,Training,Image reconstruction,Three-dimensional displays | Journal | 39 |
Issue | ISSN | Citations |
5 | 0278-0062 | 2 |
PageRank | References | Authors |
0.36 | 0 | 7 |
Name | Order | Citations | PageRank |
---|---|---|---|
Xingtong Liu | 1 | 13 | 5.02 |
Ayushi Sinha | 2 | 24 | 6.72 |
Masaru Ishii | 3 | 141 | 16.84 |
Hager Gregory D | 4 | 1946 | 159.37 |
Austin Reiter | 5 | 164 | 13.02 |
Russell H. Taylor | 6 | 1970 | 438.00 |
mathias unberath | 7 | 56 | 24.46 |