Speaker verification method based on deep neural network, terminal and storage medium

The invention discloses a speaker verification method based on a deep neural network, a terminal and a storage medium. The method comprises the following steps: acquiring audio data of a plurality of to-be-tested speakers and audio data of a plurality of registered speakers; extracting a plurality of audio features of the to-be-tested speaker and a plurality of audio features of the registered speaker; generating an embedded center vector of the extracted audio features through the trained deep neural network; wherein the deep neural network is a network constructed based on a residual neural network and a long-short term memory network; and calculating the cosine similarity of the embedded center vector between the to-be-tested speaker and the registered speaker, and verifying the to-be-tested speaker according to the cosine similarity. The method makes full use of the frequency domain feature and time domain feature information of the audio data, pays attention to the audio data which are not easy to distin