Display Technique for Embedding Information in Real Object Images Using Temporally and Spatially Luminance-Modulated Light
Display Technique for Embedding Information in Real Object Images Using Temporally and Spatially Luminance-Modulated Light
复制标题
使用时间和空间亮度调制光在真实物体图像中嵌入信息的显示技术
DOI：
10.1109/tia.2017.2726499
复制
发表时间：
2017
影响因子：
4.4
通讯作者：
Hiroshi Unno and Kazutake Uehira
中科院分区：
文献类型：
--
作者：
島田周平・上田元(編)，成澤徳子，水野一晴，遠藤聡子，池谷和信，寺谷亮司，佐川徹，松村圭一郎，佐藤廉也，藤岡悠一郎，丸山淳子，伊藤千尋，小川さやか，大門碧，遠藤貢，福西隆弘，西浦昭雄，吉田栄一，目黒紀夫，荒木美奈子，松本尚之;Satoshi Fujita;佐藤 健哉;Hiroshi Unno and Kazutake Uehira
我们提出了一种显示技术，它可以将信息隐形地嵌入到用摄像机捕获的真实对象的图像中。它使用照明灯来照亮真实的对象。这道光隐形地包含着信息。由于光包含信息，因此被这种光照射的对象的图像也包含信息。光中的信息是通过以半帧频率根据嵌入图案以小幅度调制亮度来产生的。总结了一段时间内每隔一帧图像之间的差异。通过每一帧中的调制来累加亮度的变化，同时通过该处理去除对象图像。这使得能够读出嵌入的图案。然而，对象图像不能被完全去除，并且仍然作为噪声保留。这降低了不可见模式的可读性。在本文中，我们提出了一种通过大幅降低噪声来提高不可见图案的可读性的技术。我们提出了一种利用嵌入和未嵌入模式的通道之间的相关性来降低噪声的方法。此外，它还利用嵌入图案的大小特征来降低噪声。实验结果证明了该方法的有效性。
We propose a display technique that can invisibly embed information into an image of a real object captured with a video camera. It uses illumination light that illuminates a real object. This light invisibly contains information. As the light contains information, the image of an object illuminated by such light also contains information. Information in the light is produced by modulating luminance with a small amplitude according to the embedded pattern at half-frame frequency. The differences between every other frame image over a certain period are summed up. Changes in brightness by modulation in each frame are accumulated over the frames, while the object image is removed by this processing. This makes it possible to read out the embedded patterns. However, the object image cannot be removed completely and still remains as noise. This lowers the readability of an invisible pattern. In this paper, we present a technique that improves the readability of an invisible pattern by reducing the noise drastically. We propose a method that uses the correlation between channels in which the pattern is embedded and not embedded to reduce the noise. Moreover, it also uses the size features of an embedded pattern to reduce the noise. We demonstrate the effectiveness of the proposed technique from experimental results.