藉由色彩嵌入和色彩量化實現可逆的彩色灰階影像轉換

Reversible Color to Gray Using Color Embedding and K-means Color Quantization

游采青(Tsai-Ching Yu)

指導教授 ： 貝蘇章

國立臺灣大學/電機資訊學院/電信工程學研究所/碩士(2021年)

https://doi.org/10.6342/NTU202101139

摘要

過去二十年，隨著通訊設備以及智慧型手機的普及，影像的保存以及有效率的傳輸都是我們值得探討的主題。現代人們將影像保留於移動設備中，儲存容量以及高畫質的平衡，促使著影像壓縮以及更多元化的影像演算法發展。此篇論文分別介紹彩色灰階影像轉換、可逆資料隱藏，以及可逆彩色灰階影像轉換，在各章節中我們介紹相關方法，以及使用不同的量化指標評估各個方法的實驗結果。將彩色灰階影像轉換應用於評估可逆彩色灰階影像轉換中的灰階影像，將可逆資料隱藏應用於我們提出的可逆彩色灰階影像轉換，再使用色彩量化壓縮彩色的資訊量以完整整個系統架構。提出的系統架構中，在傳送端先將彩色資訊編碼，隱藏於彩色影像的單一通道灰階影像中。而在接收端將灰階影像解碼回原灰階影像，並同時取出所藏匿的彩色資料，還原出原始三個通道的彩色影像。整體而言，與其他相關可逆彩色灰階影像轉換方法相比，時間上及綜合評量上皆有極佳的效果。

關鍵字

彩色灰階影像轉換 ； 可逆資料隱藏 ； 可逆彩色灰階影像轉換 ； 色彩量化

並列摘要

With the proliferation of communication devices and smartphones over the past two decades, the preservation and efficient transmission of images have been a topic worth exploring. The balance between storage capacity and high image quality has led to the development of image compression and more diverse image algorithms.

In this paper, we introduce color-to-gray image conversion, reversible data hiding, and reversible color-to-gray image conversion. In each chapter, we present the methods related to the topic and evaluate the experimental results of each topic using different quantitative metrics. Color-to-gray conversion is applied to evaluate gray images in reversible color-to-gray conversion. Reversible data hiding is applied to our proposed reversible color-to-gray conversion. In the proposed architecture, the color information is encoded in a gray image of the color image on the transmitter side. In addition, color quantization is performed to compress the amount of color information. At the receiving end, the gray image is decoded back to the original gray image, and the hidden color information is extracted at the same time to restore the original three-channel RGB color image. Overall, it has an excellent performance in terms of time and evaluation compared with other related methods of reversible color-to-gray conversion.

並列關鍵字

Color to gray ； Reversible data hiding ； Reversible color to gray ； Color quantization

參考文獻

[1]Qiegen Liu, Peter X. Liu, Weisi Xie, Yuhao Wang, and Dong Liang. Gcsdecolor: Gradient correlation similarity for efficient contrast preserving decolorization. IEEE Transactions on ImageProcessing, 24(9):2889–2904, 2015.[2]Qiegen Liu, Guangpu Shao, Yuhao Wang, Junbin Gao, and Henry Leung. Log­euclidean metrics for contrast preserving decolorization. IEEE Transactions on Image Processing, 26(12):5772–5783, 2017.

[3]Vishvanathan Sowmya, Divu Govind, and Kp Soman. Significance of incorporating chrominance information for effective color­-to­-grayscale image conversion. Signal, Image and Video Processing, 11, 01 2017.

[4]Wei Wang, Zhengguo Li, and Shiqian Wu. Color contrast-­preserving decolorization. IEEE Transactions on Image Processing, 27(11):5464–5474, 2018.[5]Suah Kim, Xiaochao Qu, Vasily Sachnev, and Hyoung Joong Kim. Skewed histogram shifting for reversible data hiding using a pair of extreme predictions. IEEE Transactions on Circuits and Systems for Video Technology, 29(11):3236–3246,2019.

[6]Suah Kim, Rolf Lussi, Xiaochao Qu, Fangjun Huang, and Hyoung Joong Kim. Re­versible data hiding with automatic brightness preserving contrast enhancement. IEEE Transactions on Circuits and Systems for Video Technology, 29(8):2271–2284, 2019.

[7]Ricardo L. de Queiroz and Karen M. Braun. Color to gray and back: color embedding into textured gray images. IEEE Transactions on Image Processing, 15(6):1464–1470, 2006.