RT info:eu-repo/semantics/article
T1 Gauze Detection and Segmentation in Minimally Invasive Surgery Video Using Convolutional Neural Networks
A1 Sánchez Brizuela, Guillermo
A1 Santos-Criado, Francisco-Javier
A1 Sanz-Gobernado, Daniel
A1 Fuente López, Eusebio de la
A1 Fraile Marinero, Juan Carlos
A1 Pérez Turiel, Javier
A1 Cisnal de la Rica, Ana
K1 convolutional neural networks; image segmentation; image object detection; surgical tool detection; minimally invasive surgery
AB Medical instruments detection in laparoscopic video has been carried out to increase theautonomy of surgical robots, evaluate skills or index recordings. However, it has not been extendedto surgical gauzes. Gauzes can provide valuable information to numerous tasks in the operatingroom, but the lack of an annotated dataset has hampered its research. In this article, we presenta segmentation dataset with 4003 hand-labelled frames from laparoscopic video. To prove thedataset potential, we analyzed several baselines: detection using YOLOv3, coarse segmentation, andsegmentation with a U-Net. Our results show that YOLOv3 can be executed in real time but providesa modest recall. Coarse segmentation presents satisfactory results but lacks inference speed. Finally,the U-Net baseline achieves a good speed-quality compromise running above 30 FPS while obtainingan IoU of 0.85. The accuracy reached by U-Net and its execution speed demonstrate that preciseand real-time gauze segmentation can be achieved, training convolutional neural networks on theproposed dataset.
PB MDPI
YR 2022
FD 2022
LK https://uvadoc.uva.es/handle/10324/65761
UL https://uvadoc.uva.es/handle/10324/65761
LA eng
NO Gauze Detection and Segmentation in Minimally Invasive Surgery Video Using Convolutional Neural Networks. Sensors 2022, 22, 5180
DS UVaDOC
RD 27-nov-2024