Acoustic image file tamper-proofing method

The invention relates to the technical field of deep learning, in particular to an acoustic image file tamper-proofing method, which aims at preventing the problem of an acoustic image file tamper event and specifically comprises four steps. The acoustic image file tamper-proofing method is provided based on the deep learning technology, the situation that the video file and the audio file in the sound image file are strongly correlated can be effectively solved, in addition, the video feature extraction efficiency can be effectively improved through the core frame extraction technology, finally, the acoustic image file features are solidified based on the block chain technology, and the acoustic image file tampering event can be effectively prevented. 本发明涉及深度学习技术领域，具体是一种声像文件防篡改方法，为了防止声像文件篡改事件的发生问题，具体包括四大步骤。本发明基于深度学习技术，提出了一种声像文件防篡改方法，可以有效的解决声像文件中视频文件与音频文件存在强关联的情况，此外，核心帧提取技术，可以有效的提高视频特征提取效率，最后，基于区块链技术对声像文件特征进行固化，有效的预防声像文件篡改事件的发生。