查詢結果分析
來源資料
相關文獻
- 利用聽覺遮蔽之易碎型音訊浮水印系統
- 高速無線區域網路系統通道特性與RAKE接收機設計考量
- 多重時序參考點的碼追蹤機制
- 高速無限區域網路系統通道特性與等化器設計應用考量
- 無線數據進接系統RF模組簡介
- Design and Testing of a Mixed Signal Matched Filter for IS-95 CDMA Code Acquisition
- 直接序列展頻系統之非同調碼追蹤設計
- 圖書館導入無線區域網路的探討
- 利用時基控移器設計展頻通訊系統之擬亂數碼產生器
- A Novel Differentially Coherent PN Code Acquisition Detector for DS-SS Communication Systems
頁籤選單縮合
題 名 | 利用聽覺遮蔽之易碎型音訊浮水印系統=Fragile Audio Watermarking System Using Auditory Masking |
---|---|
作 者 | 簡福榮; 郭俊榮; | 書刊名 | 臺北科技大學學報 |
卷 期 | 39:1 民95.03 |
頁 次 | 頁13-27 |
分類號 | 312.9 |
關鍵詞 | 人耳聽覺模型; 展頻通訊技術; 直接序列展頻; 二元相位移鍵法; 相位頻譜編碼法; Psychoacoustic model; Spread spectrum communication; Direct sequence spread spectrum; DSSS; Phase spectrum coding; PSC; |
語 文 | 中文(Chinese) |
中文摘要 | 拜今日網路通訊科技與數位音訊處理技術的快速發展,商業市場中的各種音樂數位音訊易於被人們非法傳輸下載或不正當的複製。本論文提出一基於人耳聽覺模型(psychoacoustic model)和展頻通訊(spread spectrum communication)的易碎型音訊浮水印技術用以檢視數位音訊完整性,確認其有否遭到竄改。此方法稱為直接序列展頻∕相位頻譜編碼法(Direct Sequence Spread Spectrum/Phase Spectrum Coding,DSSS/PSC),係將音訊浮水印嵌入於音訊訊號的頻域內,藉由改變數位音訊中小於遮蔽曲線處的相位作為浮水印的隱藏。由於嵌入到數位音訊的音訊浮水印都位在數位音訊的聽覺遮蔽門檻值(masking threshold)以下將不會影響音訊品質。實驗結果發現兩種直接序列展頻/相位頻譜編碼法DSSS/PSC-I 與DSSS/PSC-II 所取回音訊浮水印之字元正確率分別高達99.64%與100%。在經攻擊後,如MPEG-1 Layer III(MP3)壓縮攻擊、音頻位移攻擊、重新取樣攻擊、高通濾波器攻擊、低通濾波器攻擊、改變播放速度攻擊及截取一小段波形攻擊,其取回之字元正確率整體平均會遽降至0.214%及0.03571%,證實此法確具易碎性。 |
英文摘要 | With rapid development of digital audio processing and the Internet communication, digital music in market is apt to be illegally transported and duplicated by people. In this paper a fragile audio watermarking system that makes use of human auditory masking and spread spectrum communication is investigated for the purpose of content authentication even after the slightest modification. The system employs the direct sequence spread spectrum and phase spectrum coding (DSSS/PSC) to embed a digital watermark into the cover audio without sacrificing the perceptual quality. The principle is by changing the phase of the frequency spectrum that is below the auditory masking threshold of the cover audio. We analyze two slightly different schemes called DSSS/PSC-I and DSSS/PSC-II, respectively. The experimental results show that the extracted word correct rates are 99.64% for DSSS/PSC-I and 100% for DSSS/PSC-II, respectively. When suffering an attack, e.g. MP3 compression, resampling, clipping, pitch shifting, time-scale modification, and filtering, etc, the word correct rates are sharply down to 0.214% for DSSS/PSC-I and 0.03571% for DSSS/PSC-II, respectively. This proves the both are indeed good fragile watermarking schemes for verifying the completeness of the cover audio. |
本系統中英文摘要資訊取自各篇刊載內容。