Single-shot color object reconstruction through scattering medium based on neural network

Abstract The reconstruction of the color object hidden behind the scattering medium is very important because the human eye is much more sensitive to color than grayscale. Traditional methods are still difficult to reconstruct an accurate image of the hidden target from one single speckle image. In this paper, a single-shot color object reconstruction technique is proposed by designing a Color Anti-scattering Convolutional Neural Network (CASNet), which is trained to output the color and structure of the hidden color target from the input of a single speckle image. The proposed technique enables us to reconstruct the target with accurate color and structure from a broadband speckle image, and the average PSNR of recovered targets with complex structure is higher than 24dB. Efficiency and accurateness are verified through experiments.

[1]  M. Fink,et al.  Non-invasive single-shot imaging through scattering layers and around corners via speckle correlations , 2014, Nature Photonics.

[2]  Jianying Zhou,et al.  High speed color imaging through scattering media with a large field of view , 2016, Scientific Reports.

[3]  Rafael Piestun,et al.  Color image projection through a strongly scattering wall. , 2012, Optics express.

[4]  Lei Zhu,et al.  Color imaging through scattering media based on phase retrieval with triple correlation , 2019, Optics and Lasers in Engineering.

[5]  Guohai Situ,et al.  Learning-based lensless imaging through optically thick scattering media , 2019, Advanced Photonics.

[6]  J. Goodman.  Speckle Phenomena in Optics: Theory and Applications , 2020 .

[7]  Wenze Shao,et al.  A Simple and Robust Deep Convolutional Approach to Blind Image Denoising , 2019, 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW).

[8]  Michael E. Gehm,et al.  Single-shot multispectral imaging through a thin scatterer , 2019, Optica.

[9]  Wencheng Wu,et al.  The CIEDE2000 color-difference formula: Implementation notes, supplementary test data, and mathematical observations , 2005 .

[10]  Xiaohan Li,et al.  Single-shot memory-effect video , 2018, Scientific Reports.

[11]  Jing Han,et al.  Learning-based method to reconstruct complex targets through scattering medium beyond the memory effect. , 2020, Optics express.

[12]  Huichang Zhuang,et al.  Imaging of objects through a thin scattering layer using a spectrally and spatially separated reference. , 2018, Optics express.

[13]  Lei Tian,et al.  Deep speckle correlation: a deep learning approach toward scalable imaging through scattering media , 2018, Optica.

[14]  Shu-Tao Xia,et al.  Second-Order Attention Network for Single Image Super-Resolution , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Guangming Shi,et al.  Denoising Prior Driven Deep Neural Network for Image Restoration , 2018, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[16]  Wei Li,et al.  Imaging through scattering layers exceeding memory effect range by exploiting prior information , 2019, Optics Communications.

[17]  Pingxing Chen,et al.  Imaging through scattering layers exceeding memory effect range with spatial-correlation-achieved point-spread-function. , 2018, Optics letters.

[18]  Huijuan Li,et al.  Simulation and experimental verification for imaging of gray-scale objects through scattering layers. , 2016, Applied optics.

[19]  Jun Tanida,et al.  Noninvasive three-dimensional imaging through scattering media by three-dimensional speckle correlation. , 2019, Optics letters.

[20]  S. Gigan,et al.  Speckle-based hyperspectral imaging combining multiple scattering and compressive sensing in nanowire mats. , 2017, Optics letters.

[21]  George Barbastathis,et al.  Imaging through glass diffusers using densely connected convolutional networks , 2017, Optica.

[22]  Youwen Liu,et al.  Non-invasive depth-resolved imaging through scattering layers via speckle correlations and parallax , 2017 .

[23]  M. Luo,et al.  The development of the CIE 2000 Colour Difference Formula , 2001 .

[24]  Lei Zhu,et al.  Tracking moving targets behind a scattering medium via speckle correlation. , 2018, Applied optics.

[25]  Qionghai Dai,et al.  Non-invasive imaging through strongly scattering media based on speckle pattern estimation and deconvolution , 2018, Scientific Reports.

[26]  Junqi Li,et al.  Blind position detection for large field-of-view scattering imaging , 2020 .

[27]  Guihua Zeng,et al.  Image reconstruction through dynamic scattering media based on deep learning. , 2019, Optics express.

[28]  Wolfgang Osten,et al.  Tracking moving object beyond the optical memory effect , 2020 .

[29]  Guang-Can Guo,et al.  Deep hybrid scattering image learning , 2018, Journal of Physics D: Applied Physics.

[30]  Qionghai Dai,et al.  Prior-information-free single-shot scattering imaging beyond the memory effect. , 2019, Optics letters.

[31]  Arnold W. M. Smeulders,et al.  The Amsterdam Library of Object Images , 2004, International Journal of Computer Vision.