# US 2022/0036879 A1 · Method and apparatus for mining feature information, and electronic device > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2022/0036879 A1 - Document ID: US20220036879A1 - Status: Published application - Kind: A1 - Application number: 17/500,188 - Filed: 2021-10-13 - Published: 2022-02-03 - Earliest claimed: 2020-11-23 - Assignees: Beijing Baidu Netcom Science Technology Co., Ltd. - Applicants: Beijing Baidu Netcom Science Technology Co., Ltd. - Inventors: Jiaxiang GE; Zhen WU; Maoren ZHOU; Qiguang ZANG; Ming WEN; Xiaoyin FU - Web page: https://patentgrep.com/patent/US20220036879A1 ## CPC - G10L 15/02: Feature extraction for speech recognition; Selection of recognition unit - G10L 13/033: Voice editing, e.g. manipulating the voice of the synthesiser - G10L 15/063: {Training} - G10L 15/20: Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech (G10L21/02 takes precedence) - G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING ## Abstract A method for mining feature information, an apparatus for mining feature information and an electronic device are disclosed. The method includes: determining a usage scenario of a target device; obtaining raw audio data including real scenario data, speech synthesis data, recorded audio data and other media data; generating target audio data of the usage scenario by simulating the usage scenario based on the raw audio data; and obtaining feature information of the usage scenario by performing feature extraction on the target audio data. ## Claims ### Claim 1 A method for mining feature information, comprising: determining a usage scenario of a target device; obtaining raw audio data including real scenario data, speech synthesis data, recorded audio data and other media data; generating target audio data of the usage scenario by simulating the usage scenario based on the raw audio data; and obtaining feature information of the usage scenario by performing feature extraction on the target audio data. ### Claim 2 (depends on 1) The method according to claim 1, wherein generating the target audio data of the usage scenario by simulating the usage scenario based on the raw audio data comprises: obtaining scenario audio data of the usage scenario; and generating the target audio data by adding the scenario audio data to the raw audio data. ### Claim 3 (depends on 2) The method according to claim 2, wherein the scenario audio data comprises spatial reverberation data of the usage scenario, and obtaining the scenario audio data of the usage scenario comprises: obtaining attribute information of the usage scenario; obtaining state information of the target device in the usage scenario and device information of the target device; and generating the spatial reverberation data of the usage scenario based on the attribute information, the state information and the device information. ### Claim 4 (depends on 2) The method according to claim 2, wherein the scenario audio data comprises environmental noise data of the usage scenario, and obtaining the scenario audio data of the usage scenario comprises: obtaining the environmental noise data of the usage scenario from a noise database of the environmental noise data. ### Claim 5 (depends on 4) The method according to claim 4, wherein the environmental noise data comprises at least two types of environmental noise data, and obtaining the environmental noise data of the usage scenario from the noise database of the environmental noise data, comprises: obtaining a type of the usage scenario, and obtaining ratios of the at least two types of environmental noise data based on the type of the usage scenario; and obtaining the at least two types of environmental noise data randomly from their respective noise databases based on the ratios of the at least two types of environmental noise data. ### Claim 6 (depends on 2) The method according to claim 2, wherein the scenario audio data comprises residual noise data of internal noises of the target device, and obtaining the scenario audio data of the usage scenario comprises: obtaining the residual noise data of the usage scenario from a residual database of the residual noise data. ### Claim 7 (depends on 2) The method according to claim 2, wherein generating the target audio data by adding the scenario audio data to the raw audio data, comprises: selecting candidate scenario audio data from the scenario audio data for at least one time, wherein the candidate scenario audio data selected for each time is different; and adding the candidate scenario audio data selected for each time to the raw audio data. ### Claim 8 (depends on 1) The method according to claim 1, after obtaining the feature information of the usage scenario, further comprising: constructing or training a speech model of the usage scenario based on the feature information of the usage scenario. ### Claim 9 An apparatus for mining feature information, comprising: at least one processor; and a memory communicatively coupled to the at least one processor and stored with instructions executable by the at least one processor, wherein the at least one processor is configured to: determine a usage scenario of a target device; obtain raw audio data including real scenario data, speech synthesis data, recorded audio data and other media data; generate target audio data of the usage scenario by simulating the usage scenario based on the raw audio data; and obtain feature information of the usage scenario by performing feature extraction on the target audio data. ### Claim 10 (depends on 9) The apparatus according to claim 9, wherein the at least one processor is further configured to: obtain scenario audio data of the usage scenario; and generate the target audio data by adding the scenario audio data to the raw audio data. ### Claim 11 (depends on 10) The apparatus according to claim 10, wherein the scenario audio data comprises spatial reverberation data of the usage scenario, and the at least one processor is further configured to: obtain attribute information of the usage scenario; obtain state information of the target device in the usage scenario and device information of the target device; and generate the spatial reverberation data of the usage scenario based on the attribute information, the state information and the device information. ### Claim 12 (depends on 10) The apparatus according to claim 10, wherein the scenario audio data comprises environmental noise data of the usage scenario, and the at least one processor is further configured to: obtain the environmental noise data of the usage scenario from a noise database of the environmental noise data; wherein the environmental noise data comprises at least two types of environmental noise data, and the at least one processor is further configured to: obtain a type of the usage scenario, and obtain ratios of the at least two types of environmental noise data based on the type of the usage scenario; and obtain the at least two types of environmental noise data randomly from their respective noise databases based on the ratios of the at least two types of environmental noise data. ### Claim 13 (depends on 10) The apparatus according to claim 10, wherein the scenario audio data comprises residual noise data of internal noises of the target device, and the at least one processor is further configured to: obtain the residual noise data of the usage scenario from a residual database of the residual noise data. ### Claim 14 (depends on 10) The apparatus according to claim 10, wherein the at least one processor is further configured to: select candidate scenario audio data from the scenario audio data for at least one time, wherein the candidate scenario audio data selected for each time is different; and add the candidate scenario audio data selected for each time to the raw audio data. ### Claim 15 (depends on 9) The apparatus according to claim 9, wherein the at least one processor is further configured to: construct or train a speech model of the usage scenario based on the feature information of the usage scenario. ### Claim 16 A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are configured to cause a computer to execute a method for mining feature information, the method comprises: determining a usage scenario of a target device; obtaining raw audio data including real scenario data, speech synthesis data, recorded audio data and other media data; generating target audio data of the usage scenario by simulating the usage scenario based on the raw audio data; and obtaining feature information of the usage scenario by performing feature extraction on the target audio data. ### Claim 17 (depends on 16) The storage medium according to claim 16, wherein generating the target audio data of the usage scenario by simulating the usage scenario based on the raw audio data comprises: obtaining scenario audio data of the usage scenario; and generating the target audio data by adding the scenario audio data to the raw audio data. ### Claim 18 (depends on 17) The storage medium according to claim 17, wherein the scenario audio data comprises spatial reverberation data of the usage scenario, and obtaining the scenario audio data of the usage scenario comprises: obtaining attribute information of the usage scenario; obtaining state information of the target device in the usage scenario and device information of the target device; and generating the spatial reverberation data of the usage scenario based on the attribute information, the state information and the device information. ### Claim 19 (depends on 17) The storage medium according to claim 17, wherein the scenario audio data comprises environmental noise data of the usage scenario, and obtaining the scenario audio data of the usage scenario comprises: obtaining the environmental noise data of the usage scenario from a noise database of the environmental noise data; wherein the environmental noise data comprises at least two types of environmental noise data, and obtaining the environmental noise data of the usage scenario from the noise database of the environmental noise data, comprises: obtaining a type of the usage scenario, and obtaining ratios of the at least two types of environmental noise data based on the type of the usage scenario; and obtaining the at least two types of environmental noise data randomly from their respective noise databases based on the ratios of the at least two types of environmental noise data. ### Claim 20 (depends on 17) The storage medium according to claim 17, wherein the scenario audio data comprises residual noise data of internal noises of the target device, and obtaining the scenario audio data of the usage scenario comprises: obtaining the residual noise data of the usage scenario from a residual database of the residual noise data.