音訊秘密分享以數學計算的方式進行機密音訊的分享與還原,故而能提供良好的還原品質,甚至以無失真還原的方式出現。音訊秘密分享技術具有(t,n)門檻值(t<=n 且 2<=n)的特性。分享時將機密音訊分享成n把金鑰及與金鑰相對應的分享音訊,n為分享者的數目,且每位分享者僅握有一把金鑰及一份分享音訊,還原時只要取得t把金鑰及與金鑰相對應的分享音訊,經過還原計算後,即可獲得原機密音訊。如此,我們便無須擔心部份金鑰及分享音訊遺失會造成機密音訊無法復原,或是因複製多份機密音訊而降低安全性的問題。本論文採用極具效益的中國餘式定理為分享技術,在Android平台實際開發以達到高效率的音訊秘密分享系統。
This study presents an (s, t, n) progressive method for sharing an audio. In an (s, t, n) progressive audio sharing scheme, n shared audios are generated from the secret audio, gathering s shared audios acquires coarse resolution of secret audio, and using t shared audios can reconstruct the original secret audio losslessly. The Chinese Remainder Theorem is adopted in the proposed scheme to share coefficients acquired from 1-D integer wavelet transform with different thresholds. First, the maximum wavelet level number needed in 1-D integer wavelet transform is determined from thresholds s and t. Then, the proposed scheme applies the secret audio to 1-D integer wavelet transform for acquiring coefficients under different levels. At last, all wavelet coefficients are partitioned to (t-s+1) groups and share each group with different thresholds under inverse wavelet transform. Experimental results demonstrate that the proposed scheme can share secret audios efficiently and progressively.