ICA在监听键盘录入信息中的应用

Application of ICA in Intercepting Messages by Sounds of Keystrokes

文中针对在强环境噪声背景下,利用键盘敲击音对输入内容进行判定的精度受到很大影响的现象,通过对采集到的强噪背景下键盘敲击音进行盲分离,提取出清晰的键盘敲击音,从而提高判定精度。对实测数据进行盲分离的结果表明,快速独立分量分析方法是一种有效的盲信号提取方法,此方法可分离出湮没在强噪声环境中的键盘敲击音,具有很高的可靠性。

Because the determinant precision is greatly influenced by the typing sound in the background of strong noises, it is very important to extract the clear typing sound of by blind source separation and improve the determinant precision. The blind source separation of the practically measured data show that the FastICA is a effective method for the extraction of blind signal, which could effectively separate the typing sound sunk in the strong noisy environment. This method is of practicability and reliability.