講演情報
[20a-A305-2]Deep learning based high-speed three-dimensional dynamic optical coherence tomography generation for tumor spheroid evaluation
〇(M2)Yusong Liu1, Ibrahim Abd El-Sadek1,2, Shuichi Makita1, Tomoko Mori3, Atsuko Furukawa3, Satoshi Matsusaka3, Yoshiaki Yasuno1 (1.COG, Univ. of Tsukuba, 2.Faculty of Science, Damietta Univ., 3.Faculty of Medicine, Univ. of Tsukuba)
キーワード：
Optical coherence tomography、tumor spheroid、deep learning
We proposed a deep-learning-based method that requires short acquisition time to achieve high-speed three-dimensional (3-D) dynamic optical coherence tomography (D-OCT) for tumor spheroid evaluation. This method requires volumetric acquisition time of 6.55 seconds, which is only 12.5% of that of conventional 3-D D-OCT algorithm. D-OCT generated by this method shows similar image appearance and numerical metrics to conventional 3-D D-OCT for tumor spheroids. Our NN-based LIV algorithm has potential for high-speed three-dimensional D-OCT generation and replacing conventional 3-D D-OCT for 3-D evaluation of tumor spheroid.