Abstract: The present disclosure relates to a method and device for selecting from a plurality of beams. The method includes: obtaining a plurality of beam data, and performing frequency sampling on each of the plurality of beam data; obtaining a plurality of beam frequency correlation coefficients based on frequency sampling data of each of the plurality of beam data, in which a beam frequency correlation coefficient is configured to indicate a similarity between one in the plurality of beam data and another one in the plurality of beam data; obtaining a beam frequency correlation coefficient sum corresponding to each of the plurality of beam data based on the plurality of beam frequency correlation coefficients; and selecting beam data having the beam frequency correlation coefficient sum satisfying a preset correlation coefficient requirement in the plurality of beam data as target beam data.

Abstract: A method for responding to a voice signal includes: receiving a voice signal by the plurality of electronic devices; determining, for each of the electronic devices, a target sound source point associated with the voice signal; determining, for each of the electronic devices, a direct sound energy associated with the voice signal according to the target sound source point; and selecting from the plurality of electronic devices at least one electronic device with a direct sound energy satisfying a predetermined condition, and responding to the voice signal by the at least one electronic device. A sound pickup decision system based on a plurality of electronic devices, can make more accurate, more reasonable, and more targeted response to a user's voice, improving voice interaction experience.

Abstract: A method for processing an audio signal is provided. In the method, audio signals sent by at least two sound sources are acquired by at least two microphones to obtain multiple frames of original noisy signals of each microphone on a time domain. For each frame, frequency-domain estimation signals of each sound source are acquired according to the original noisy signals of the at least two microphones. For each sound source, the frequency-domain estimation signals are divided into multiple frequency-domain estimation components on a frequency domain. For each sound source, feature decomposition is performed on a related matrix of each frequency-domain estimation component to obtain a target feature vector. A separation matrix of each frequency point is obtained based on target feature vectors and the frequency-domain estimation signals. The audio signals of sounds are obtained based on the separation matrixes and the original noisy signals.

Abstract: A space division method includes: receiving a first sound signal that is a medium-high frequency sound signal; decoding the first sound signal by specified decoding to obtain device information of a sound source device that emits the first sound signal; and generating space division information when the device information of the sound source device is successfully obtained, wherein the space division information is configured to indicate that the sound source device and the sound collection device are located in the same spatial region.

Abstract: A facial image enhancement method includes: determining a plurality of feature regions in a facial image according to feature points of the facial image; correcting an initial enhancement weight coefficient of each feature region of the feature regions and obtaining a weight map of discrete enhancement on the facial image; obtaining a weight map of continuous enhancement on the facial image according to the weight map of discrete enhancement on the facial image; and performing enhancing the facial image according to the weight map of continuous enhancement on the facial image to obtain an enhanced facial image.

Abstract: A facial image enhancement method includes: determining a plurality of feature regions in a facial image according to feature points of the facial image; correcting an initial enhancement weight coefficient of each feature region of the feature regions and obtaining a weight map of discrete enhancement on the facial image; obtaining a weight map of continuous enhancement on the facial image according to the weight map of discrete enhancement on the facial image; and performing enhancing the facial image according to the weight map of continuous enhancement on the facial image to obtain an enhanced facial image.

Abstract: A method for training an image processing model, includes: acquiring M frames of long-distance shot images of each of N training scenes, both N and M being positive integers greater than 1; acquiring a short-distance standard image of each training scene; for each training scene, clipping the M frames of long-distance shot images to obtain a same field of view as that of the short-distance standard image, and aligning the M frames of long-distance shot images after being clipped with the short-distance standard image; forming a training sample couple of each training scene by the M frames of long-distance shot images and the short-distance standard image after alignment, and forming a training sample set by N training sample couples corresponding to the N training scenes; and training an image processing model based on the training sample set.

Abstract: Provided are an audio signal processing method and device. The method can include acquiring audio signals from at least two sound sources through at least two microphones to obtain multiple frames of original noise signals of the at least two microphones in a time domain, and, for each frame in the time domain, acquiring respective frequency-domain estimated signals of the at least two sound sources according to the respective original noise signals. The method can further include, for each sound source, dividing the frequency-domain estimated signal into frequency-domain estimated components which each corresponds to a frequency-domain sub-band and includes multiple frequency point data in a frequency domain, determining a weighting coefficient of each frequency point in the frequency-domain sub-band, and updating a separation matrix of each frequency point according to the weighting coefficient and obtaining the audio signals based on the updated separation matrices and the original noise signals.

Abstract: A method for processing audio signal includes that: audio signals emitted respectively from at least two sound sources are acquired through at least two microphones to obtain respective original noisy signals of the at least two microphones; sound source separation is performed on the respective original noisy signals of the at least two microphones to obtain respective time-frequency estimated signals of the at least two sound sources; a mask value of the time-frequency estimated signal of each sound source in the original noisy signal of each microphone is determined based on the respective time-frequency estimated signals; the respective time-frequency estimated signals of the at least two sound sources are updated based on the respective original noisy signals of the at least two microphones and the mask values; and the audio signals emitted respectively from the at least two sound sources are determined.

Abstract: The disclosure relates to a method, an apparatus, a system and a device for controlling an intelligent device, and a storage medium. The method includes collecting voice data of a user; performing semantic recognition processing on the collected voice data of the user to obtain a recognition result; determining a target intelligent device based on device information in the recognition result; generating a control command based on operation instruction information in the recognition result; and sending the control command to the target intelligent device.

Abstract: The present disclosure, belonging to the technical field of natural language processing, provides a text sequence segmentation method. The method can include acquiring n segmentation sub-results of the text sequence, the n segmentation sub-results being acquired by segmenting the text sequence by n segmentation models, and processing the n segmentation sub-results by a probability determination model branch in a result combination model to acquire a segmentation probability of the each segmentation position. The method can further include processing the segmentation probability of the each segmentation position by an activation function in the result combination model to acquire a segmentation result of the text sequence. According to the present disclosure, using each segmentation position in the text sequence as a unit, segmentation results of a plurality of segmentation models are combined, such that the accuracy of segmentation of a new text can be improved.

Abstract: A method for object tracking includes: obtaining frames with a number of N of history images of the object; acquiring first predicted feature point information of each frame image by using first network models corresponding to each frame image in the frames with a number of N of history images, and acquiring second predicted feature point information of the each frame image by using second network models corresponding to each frame image; adjusting parameters of the first network model and parameters of the second network model based on the first predicted feature point information and the second predicted feature point information until the first network model and the second network model are trained completely; and performing tracking of the object by using the first completely trained network model and the second completely trained network model.

Abstract: A method for object tracking includes: obtaining frames with a number of N of history images of the object; acquiring first predicted feature point information of each frame image by using first network models corresponding to each frame image in the frames with a number of N of history images, and acquiring second predicted feature point information of the each frame image by using second network models corresponding to each frame image; adjusting parameters of the first network model and parameters of the second network model based on the first predicted feature point information and the second predicted feature point information until the first network model and the second network model are trained completely; and performing tracking of the object by using the first completely trained network model and the second completely trained network model.

Abstract: A method for processing an audio signal is provided. In the method, audio signals sent by at least two sound sources are acquired by at least two microphones to obtain multiple frames of original noisy signals of each microphone on a time domain. For each frame, frequency-domain estimation signals of each sound source are acquired according to the original noisy signals of the at least two microphones. For each sound source, the frequency-domain estimation signals are divided into multiple frequency-domain estimation components on a frequency domain. For each sound source, feature decomposition is performed on a related matrix of each frequency-domain estimation component to obtain a target feature vector. A separation matrix of each frequency point is obtained based on target feature vectors and the frequency-domain estimation signals. The audio signals of sounds are obtained based on the separation matrixes and the original noisy signals.

Abstract: Provided are an audio signal processing method and device. The method can include acquiring audio signals from at least two sound sources through at least two microphones to obtain multiple frames of original noise signals of the at least two microphones in a time domain, and, for each frame in the time domain, acquiring respective frequency-domain estimated signals of the at least two sound sources according to the respective original noise signals. The method can further include, for each sound source, dividing the frequency-domain estimated signal into frequency-domain estimated components which each corresponds to a frequency-domain sub-band and includes multiple frequency point data in a frequency domain, determining a weighting coefficient of each frequency point in the frequency-domain sub-band, and updating a separation matrix of each frequency point according to the weighting coefficient and obtaining the audio signals based on the updated separation matrices and the original noise signals.

Abstract: A method for processing audio signal includes that: audio signals emitted respectively from at least two sound sources are acquired through at least two microphones to obtain respective original noisy signals of the at least two microphones; sound source separation is performed on the respective original noisy signals of the at least two microphones to obtain respective time-frequency estimated signals of the at least two sound sources; a mask value of the time-frequency estimated signal of each sound source in the original noisy signal of each microphone is determined based on the respective time-frequency estimated signals; the respective time-frequency estimated signals of the at least two sound sources are updated based on the respective original noisy signals of the at least two microphones and the mask values; and the audio signals emitted respectively from the at least two sound sources are determined.

Abstract: The present disclosure relates to a method and a terminal for performing word segmentation on text information, and a storage medium. The method includes: acquiring the text information and configuration information, in which the configuration information includes at least two first word segmentation rules; converting the first word segmentation rules into second word segmentation rules according to a predetermined rule; in response to determining that an intersection exists between character strings of the text information matched by two of the second word segmentation rules, determining that two first word segmentation rules corresponding to the two of the second word segmentation rules associated with the intersection conflict; and processing the text information according to the configuration information, and outputting a result of the word segmentation on the text information.

Abstract: A space division method includes: receiving a first sound signal that is a medium-high frequency sound signal; decoding the first sound signal by specified decoding to obtain device information of a sound source device that emits the first sound signal; and generating space division information when the device information of the sound source device is successfully obtained, wherein the space division information is configured to indicate that the sound source device and the sound collection device are located in the same spatial region.

Abstract: A method for detecting small obstacles includes: acquiring a first 3D point cloud corresponding to image data acquired by a cleaning robot; extracting, from the first 3D point cloud, a second 3D point cloud of a ground region; extracting, from the second 3D point cloud, a third 3D point cloud having a height value in a set height range; calculating a ground projection point cloud of the third 3D point cloud; and, determining morphologically-connected regions in the ground projection point cloud, and using a morphologically-connected region having an area less than a preset value as a region where a small obstacle is located. By effectively recognizing the ground, acquiring the ground projection point cloud and processing the point cloud by image processing, accuracy and timeliness of the algorithm can be improved compared with directly processing discrete 3D point clouds.

Abstract: An annotation method and device and a storage medium are provided. The annotation method includes operations as follows. A first probability value that a first sample image is annotated with an Nth tag when the first sample image is annotated with an Mth tag is determined based on first tag information of a first image set. M and N are unequal and are positive integers. The first probability value is added to second tag information of a second sample image annotated with the Mth tag in a second image set.