Speech recognition method and apparatus based on speaker recognition

An apparatus for speech recognition, the apparatus comprising:

• a receiver configured to receive a first audio signal while a session for receiving an input of one or more audio signals is maintained when the session is activated based on a user input;
• at least one processor;
• at least one memory storing one or more computer programs configured to executed by the at least one processor, wherein the one or more computer programs include instructions to at least:
• detect a first speech signal from the first audio signal,
• perform first speech recognition on the first speech signal,
• determine whether a speaker of the first speech signal is a registered speaker based on speaker information obtained from the first speech signal,
• determine whether to maintain the session based on a result of the determination of whether the speaker of the first speech signal is a registered speaker, and
• output a result of the first speech recognition,
• receive a second audio signal while the session is maintained,
• perform second speech recognition on a second speech signal detected from the second audio signal while the session is maintained,
• obtain a correlation between a meaning of the first speech signal and a meaning of the second speech signal of the second speech signal;
• and
• an output unit configured to output a result of the second speech recognition based on the obtained correlation,
• wherein an operation of the receiving the input of an audio signal is performed according to whether the session is currently maintained.

A non-transitory, computer-readable recording medium having recorded thereon at least one program comprising instructions that, when executed by at least one processor, configure the instructions to execute a speech recognition method by:

• activating, by an electronic device, a session for receiving a first audio signal based on a user input;
• performing, by the electronic device, first speech recognition on a first speech signal detected from the first audio signal while the session is maintained;
• determining, by the electronic device, whether a speaker of the first speech signal is a registered speaker based on speaker information obtained from the first speech signal;
• determining, by the electronic device, whether to maintain the session based on a result of the determination of whether the speaker of the first speech signal is a registered speaker;
• outputting, by the electronic device, a result of the first speech recognition;
• receiving, by the electronic device, a second audio signal while the session is maintained;
• performing, by the electronic device, second speech recognition on a second speech signal detected from the second audio signal while the session is maintained;
• obtaining, by the electronic device, a correlation between a meaning of the first speech signal and a meaning of the second speech signal;
• and
• outputting, by the electronic device, a result of the second speech recognition based on the obtained correlation,
• wherein an operation of the receiving the input of an audio signal is performed according to whether the session is currently activated.