基於時域能量熵最佳化的音訊隱寫與壓縮整合系統

私密訊息或對話在網路中傳輸需要受到保護，例如個人隱私、病患資訊、病患與醫護人員的某些對話等等。基於此，本論文提出一種基於能量熵（energy entropy）最佳化的時域音訊隱寫與門檻值式壓縮之整合系統。在音訊隱寫術中，訊噪比（signal-to-noise ratio, SNR）與位元錯誤率（bit error ratio, BER）之間存在著權衡，而這兩者通常作為評估音訊隱寫術之保真度、穩健性與品質的效能指標。因此，本論文發展一種基於最佳化的隱寫術。在該技術中，SNR被設為效能指標，而能量熵嵌入式則作為最佳化過程中的約束條件。同時，為了降低網路傳輸中的資料量，我們也進行了壓縮處理。此外，本論文採用人工智慧演算法中的粒子群最佳化（particle swarm optimization, PSO）來求解最佳化問題。實驗結果證實，本方法所嵌入的資訊在不論是否面對攻擊（如振幅修改與MP3壓縮），仍能維持足夠的 SNR 與低 BER，展現出良好的隱蔽性與穩健性。

Private messages or conversations need to be protected when transmitted over the Internet, such as personal privacy, patient information, and certain conversations between patients and medical staff. Therefore, we propose a time-domain audio steganography and threhosld-based compression integration system based on energy-entropy optimization. In audio steganography, there is a trade-off between signal-to-noise ratio (SNR) and bit error ratio (BER), which are usually used as performance indicators to evaluate the fidelity, robustness, and quality of audio steganography technology. Therefore, we develop a steganography method based on optimization. In this method, SNR is set as the performance indicator, and energy-entropy embedding is used as a constraint in the optimization process. At the same time, in order to reduce the amount of data in network transmission, we also perform compression processing. In addition, we use particle swarm optimization (PSO) in artificial intelligence algorithms to solve the optimization problem. Experimental results confirm that the information embedded by this method can maintain sufficient SNR and low BER regardless of attacks (such as amplitude modification and MP3 compression), showing good concealment and robustness.