以回聲資訊隱藏為基礎的音訊浮水印技術設計

Abstract

近年來，隨著資訊加速傳播，智慧財產權的保護日益受到重視，對數位形式產品之版權尤其看重。為了保護數位資訊，基於傳統所使用之在文書或圖像中的浮水印技術，衍生出多樣的數位浮水印的技術。現在數位浮水印技術已經逐漸被廣泛應用於保護數位化的多媒體資訊或產品中。在多媒體資訊中，有關音訊的部分，已有數種可適應不同形式音訊或依循人類聽覺特性設計的浮水印技術被提出，其中之一以模擬自然界回聲的方式來隱藏訊息，稱作回聲資訊隱藏。我們選擇回聲資訊隱藏技術來做進一步研究的原因在於：它與其他音訊浮水印方法比較，擁有較佳的浮水印強韌性、可接受的資訊傳輸率以及在不同聲音格式上的通用性。雖然回聲資訊隱藏技術有許多長處，但與某些音訊浮水印技術比較，其資訊傳輸率略差，並且其易受限於音訊本身特性，造成浮水印解回錯誤。在這篇論文之中，我們分析了回聲資訊隱藏系統的各項控制參數，並提出解決有關回聲資訊隱藏技術問題之方法。從實驗證明，本論文所提出之方法可改進系統的資訊傳輸量與浮水印解回正確率。

The protection of intellectual property has become more and more important in recent years, especially to products in digital formats. The digital products are protected based on adding watermark, which is traditionally used for the intellectual property protection. The watermark is embedded into digital audio signals, which is referred as a digital audio watermarking technique, based on the characteristics of audio signals or the human auditory system. One of these techniques embeds the watermark by imitating environmental echoes, which is called as an echo data hiding technique. As compared to the other digital audio watermarking techniques, the echo data hiding technique has the following features, which are better robustness, an acceptable information transmission rate, and no limitation on the type of audio formats. However, in the echo data hiding technique, the information transmission rate may be lower than some of the other approaches, and the extractability of the echo may be constrained by the characteristics of the host audio. In this thesis, we have tried to solve these problems as well as to improve the performance of the echo data hiding system.