Object recognition through scattering media using convolutional neural network

Yulin Wu; Huimin Yan

doi:10.1117/12.2532238

17 May 2019 Object recognition through scattering media using convolutional neural network

Yulin Wu, Huimin Yan

Proceedings Volume 11170, 14th National Conference on Laser Technology and Optoelectronics (LTO 2019); 111700J (2019) https://doi.org/10.1117/12.2532238
Event: Fourteenth National Conference on Laser Technology and Optoelectronics, 2019, Shanghai, China

Abstract

When coherent light propagates through scattering media, it is scattered by the uneven particles in scattering media, resulting in a seemingly random speckle pattern. However, researchers have demonstrated that the information of the object is still preserved in the speckle pattern. In this paper, the convolutional neural network (CNN), which has a powerful ability to automatically extract nonlinear features, was used to recognize phase objects displayed on the spatial light modulator through scattering media. The experimental results showed that the CNN architecture can recognize faces images hidden by ground glass with a recognition accuracy of 97%. In addition to applying CNN to binary classification problems, we also verified that CNN could be effectively applied to multi-classification tasks, and the classification accuracy of handwritten digital data sets reached 95%.

Citation Download Citation

Yulin Wu and Huimin Yan "Object recognition through scattering media using convolutional neural network", Proc. SPIE 11170, 14th National Conference on Laser Technology and Optoelectronics (LTO 2019), 111700J (17 May 2019); https://doi.org/10.1117/12.2532238

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available