多張影像的低儲存快速解碼分享與比例顯像分享、被壓縮影像在頻率域的驗證與修復、多功能的混合型視覺加密分享

Abstract

本計畫為期三年，其目的是提出影像分享、修復及視覺加密技術。本年(第一年)的主題是多張影像的「低儲存空間快速解碼型」分享與「比例顯像型」分享。在第一主題中，我們提出以下系統：在單一編碼過程中分享掉多張影像，要求具有以下兩個時空並進的優點: (1).空間優勢：影像總輸入的大小與影像總輸出的大小比例（I/O size ratio）提高為1，亦即 1:1；(2).時間優勢：重建每個像素都只要常數次的運算，亦即O(1) ，因而不受影像張數影響。第一年第二個主題是多張影像的比例顯像型分享。在本主題中，我們希望設計出一種漸進式解碼方式，當取得t1份的JPEG偽裝碼時，可以解回第1群的所有影像。當取得t2（＞t1）份的JPEG偽裝碼時，可以解回第1群和第2群的所有影像。我們提出的多張影像分享法，同時擁有漸進式解碼、無失真還原(當收集到大部份的偽裝碼時)、以及儲存空間小的優點，所以非常適合將分享後之分存隱藏至偽裝媒體(stego media)，不會對要被藏入資料的媒體影像品質有顯著的影響。

This project is the first part of a three-year project. In Year 1, we deal with the following two topics: Topic A: design a space-saving fast-decoding sharing method for multiple images. This lossless method has two advantages simultaneously: (i) space: I/O size ratio is as high as 1:1；(ii) speed: reconstructing a pixel only needs a constant number of operations； and the number is independent of the number of secret images. (No reported lossless sharing method owns these two advantages simultaneously.) Topic B: design a multiple-images sharing method which has image by image security-dependent progressive decoding, lossless reconstruction of secret (when most of the shadows are collected), and compact size of shadows. An application of this topic is that, in a meeting, if some people are late, once the number of attendance has reached a threshold, some less important images can be decoded and discussed. When more and more people attend the meeting, the more important images can be decoded and discussed.