Presentation Information
[19p-C31-4]Moving Average of Plume Movie in Thin Film Deposition Process Conditions Based on Plume Images
〇(M1)Tatsuma Hachiya1, Horio Keiichi1,5, Yamazaki Shuntaro2, Ichino Yusuke3,5, Ichinose Ataru4,5, Horide Tomoya2,5, Matsumoto Kaname2,5, Yoshida Yutaka2,5 (1.Kyushu Inst. of Technol., 2.Nagoya Univ., 3.Aichi Inst. of Technol., 4.CRIEPI, 5.JST-CREST)
Keywords:
Machine Learning,Plume,Convolutional Neural Network
We aim to estimate process parameters from plume images to improve the reproducibility of thin film deposition and to stabilize the process.
The laser power and oxygen partial pressure are changed, the plume is photographed under each condition, and parameters are estimated using a convolutional neural network that uses the plume images as input.
In this research, we consider that the variation caused by the fluctuation of the plume causes a decrease in accuracy, and we apply a moving average to the plume video image to reduce the effect of the fluctuation.
The laser power and oxygen partial pressure are changed, the plume is photographed under each condition, and parameters are estimated using a convolutional neural network that uses the plume images as input.
In this research, we consider that the variation caused by the fluctuation of the plume causes a decrease in accuracy, and we apply a moving average to the plume video image to reduce the effect of the fluctuation.
Comment
To browse or post comments, you must log in.Log in