可供普及溝通使用之「富含訊息視訊」之建構與應用

Abstract

在日常生活中，存於環境周遭的許多物體都可用來容納資訊，達到普及溝通(pervasive communication)的目的。而這些物體廣義來說均可視為多媒體，會以不同的形式出現在生活周遭，如廣告看板、海報、在電視螢幕上所放映的視訊等等。訊息若可藉由資訊隱藏的技術嵌入在這些物體之中，而隱藏的訊息亦可被人們用訊息讀取器(如：手機、平版電腦)抽取出來，則此種嵌入訊息的多媒體即稱為「富含訊息多媒體(message-rich multimedia)」。 本研究嘗試設計一新方法來製作一種新型的視訊，可不必透過直接傳輸數位檔案的方式來交換訊息。此為一新型的富含訊息多媒體，稱之為「富含訊息視訊」，可藉由資訊隱藏的方式實現；而被嵌入在視訊之中的訊息可以藉由「放映與再取像(display-and-imaging) 」的方式抽取出來。為了達到這個目標，本論文設計了一個訊息嵌入程序和一個對應的訊息抽取程序。 在訊息嵌入程序中，本研究採用了嵌入「對角形圓訊號」在視訊畫面頻譜中(video-frame spectrum)之方式，讓視訊中的每一畫面皆可藏入一個二位元訊號。另一方面，在放映影片進行再次取像時，所翻拍之影片幾乎不可能與原有影片有相同的起始點。此外，視訊有許多不同格式也必須考慮。為了解決以上的議題，本論文提出一藉由特殊設計之起始及結束信號及一與影片格式特性符合之同步機制，讓訊息嵌入程序與訊息抽取程序之間不需傳輸訊息即可達成同步。 在訊息抽取程序中，本論文提出一藉由「矩量保持門檻化(moment-preserving thresholding)」技術與一使用特殊遮罩之訊號偵測方式，讓訊息抽取程序可自適性的偵測每一畫面中的訊號，並達到很高的辨識率。接著，本論文設計了一線性資料結構，用以重現被嵌入於視訊中的訊息，此結構可使訊號更易於分析與轉換，使得所提抽取程序可成功的抽取出被嵌入的訊號。最後，本論文提出視訊取樣與畫面縮小之方式，來加速訊息抽取程序，來提高所提方法的實用性。 以上所提出系統為一創新之作，深入的理論分析及實驗結果顯示該系統具有可行性及實用性。

In people’s daily life, many kinds of identities existing in the environment can be utilized to accommodate information for the purpose of pervasive communication. These identities may all be, in a broad sense, regarded as multimedia. They can appear in our daily life in many forms like hard copies, advertisements, posters, displays on monitors or TVs, etc. Messages may be injected into such identities by data hiding techniques. Such messages can be extracted by people using message readers such as smart phones, tablets. Such multimedia with messages embedded are called message-rich multimedia. Specifically, it is tried in this study to design a new technique for creating a new type of video, by which, differently from traditional ways of information transmission, people can exchange information without sending directly digital files such as videos and images. That is, a new type of message-rich multimedia, called message-rich video, implemented by hiding messages in videos is proposed. Moreover, the message embedded in the video can be extracted in a display-and-imaging fashion. To achieve this purpose, a message embedding process and a corresponding message extraction process are designed. For the message embedding process, the use of diagonal circular signals in the frequency domain is proposed, by which the proposed system can embed a 2-bit signal in each video frame. Next, it is almost impossible to conduct imaging of a displayed video exactly from its starting point. Furthermore, videos have various formats which must be considered as well in the design of the embedding process. To overcome these issues, a synchronization mechanism by uses of certain specially-designed initial and ending signals based on different video formats is proposed, by which the system can synchronize the message extraction work with the embedded message signals without transmitting any information. For the message extraction process, a signal detection method based on the technique of moment-preserving thresholding using certain specially-designed masks is proposed, by which the system can detect signals adaptively with a high recognition rate. Also, a linear structure is designed for representing the message signals by which signals can be extracted, analyzed, and converted into ASCII codes. Furthermore, the extraction process is speeded up by frame resizing and down-sampling techniques. The feasibility and effectiveness of all the proposed methods are demonstrated by theoretical analyses and good experimental results.