146 results about "End point detection" patented technology

The invention relates to the field of voice signal processing, and discloses a multi-level labeling method and device in a dialogue scene. Through the endpoint detection and automatic speech recognition method of the dialogue audio data, noise and mute in real speech can be removed, and effective speech segments can be obtained. And carrying out clustering analysis based on the effective voice segments, obtaining target voice segments and text data of the corresponding speakers, and carrying out emotion feature extraction through multi-modal emotion analysis. And finally, realizing automatic annotation of the inquiry strategy or intention understanding of the corresponding speaker by utilizing a large language model. According to the method, the recorded conversation voice can be subjected to multi-level labeling, an efficient and accurate information management and analysis tool is provided for a conversation scene, and the arrangement efficiency of conversation records can be greatly improved. The annotation information can be used for training and improving an artificial intelligence system, so that the artificial intelligence system plays a greater role in future psychological consultation service, and the overall quality and efficiency of psychological health service are improved.

A system and method for improved endpoint detection and response (EDR) in a cloud computing environment initiates inspection based on data received from a sensor deployed on a workload. The method includes: configuring a resource, deployed in a cloud computing environment, to deploy thereon a sensor, the sensor configured to detect runtime data; detecting a potential cybersecurity threat on the resource based on detected runtime data received from the sensor; and initiating inspection of the resource for the potential cybersecurity threat.

Techniques are described herein for a method of obtaining a token based on a conversation in real time. The method further includes predicting, using a large language model (LLM) and the token, a next token. The method further includes predicting, using a classifier and the next token, a completion of a user turn. The method further includes triggering a next turn of the conversation in real time using the completion of the user turn.

The embodiment of the invention discloses an audio processing method and device, and the method comprises the steps: carrying out the traversal of a segmented sentence list which is determined through the interception of an audio stream based on a mute detection result, and determining a current segmented sentence and an endpoint detection result of the current segmented sentence, determining a start frame of the current segmented sentence as a start frame of a latest to-be-identified sentence when the current segmented sentence has a sentence start point, determining an end frame of the current segmented sentence as an end frame of the latest to-be-identified sentence when the current segmented sentence has a sentence end point, and determining the latest to-be-identified sentence according to the start frame and the end frame of the latest to-be-identified sentence. According to the method and the device, the segmented sentences in the audio stream can be continuously updated, the to-be-recognized sentences can be determined according to the endpoint detection results of the segmented sentences, continuous long speech recognition is realized, and the complete content of the to-be-recognized sentences from the start frame to the end frame is recognized, so that the speech recognition efficiency is improved. The accuracy of long speech recognition can be improved.

Methods and systems for managing an endpoint detection and response (EDR) framework for data processing systems are disclosed. To monitor for malicious entities, a management controller of a data processing system may obtain a data package including data managed by hardware resources of the data processing system. The data package may be provided to a trusted management system via an out of band communication channel and the management controller may monitor for responses from the trusted management system. If the management controller identifies a response indicating that the data processing system is in a compromised state, one or more actions may be performed to remediate the compromised state of the data processing system.

The invention relates to the technical field of single-phase disconnection detection of an ungrounded system, in particular to a single-phase disconnection fault detection method and system based on wavelet transform. A plurality of detection points are arranged on a distribution line of the power distribution network; line voltage amplitude signals of all the detection points are collected, and a data set is formed; performing wavelet transform processing on the data set to obtain a detail coefficient corresponding to each detection point; comparing the absolute value of the detail coefficient with a preset threshold value to judge a broken line fault; and determining the position of the fault section according to the detail coefficients exceeding the preset threshold. The detection points comprise a main line detection point, a branch line detection point and an end point detection point, and the detection point spacing distance is set according to the line importance degree and load distribution; and performing signal processing by adopting Haar wavelet transform to obtain a threshold value based on a detail coefficient in a normal operation state. Through a method of combining wavelet transform and adaptive threshold determination, quick and accurate detection of a broken line fault is realized without traversing a network topology structure, and the calculation complexity is greatly reduced.

The invention relates to the technical field of voice recognition intelligent home control, and discloses a non-specific person voice recognition intelligent switch control method and system based on deep learning. The non-specific person voice recognition intelligent switch control method is applied to intelligent switch control equipment, and specifically comprises the following steps of S101, receiving original audio signals continuously collected in a to-be-controlled environment, and preprocessing the collected original audio signals, and then a starting point and an ending point of an effective voice segment are positioned by adopting endpoint detection based on a double-threshold method and combining the characteristic parameters of the short-time energy and the short-time zero-crossing rate. A multi-layer hidden layer structure with Dropout regularization is adopted in a neural network model, the generalization ability of the model is enhanced, a context sensing mechanism is introduced into a semantic understanding module, a composite instruction containing azimuth information can be intelligently analyzed, crossing from recognition to understanding is achieved, and the method has the advantages of being high in practicability and easy to popularize. The system is ensured to maintain a high recognition rate for voice instructions of different users under different environment conditions.

The invention discloses a caller identity identification method and system based on dispatch call voiceprint recognition, and the method comprises the steps: collecting voice data, achieving the preprocessing of voice through voice denoising processing, voice endpoint detection and voice speed normalization processing, carrying out the voiceprint feature calculation and extraction of the preprocessed voice data, and carrying out the recognition of the voiceprint feature of a caller. Carrying out the modeling according to the voice voiceprint features, completing the building of a voiceprint feature library, carrying out the matching of the extracted voiceprint features with the voiceprint feature library, judging the identity of a caller, and completing the identity authentication of the caller; the method can effectively prevent identity fraudulent use and fraudulent behaviors, ensures the accuracy and safety of the scheduling instruction, reduces the steps of manual verification, and improves the scheduling efficiency.

Methods and systems for process chamber qualification for maintenance process endpoint detection are provided. Data collected by one or more sensors of a process chamber of a manufacturing system is identified. The identified data is collected during performance of initial maintenance operation(s) of a maintenance process. A current state of the process chamber is determined, based on the identified data, after the performance of the initial maintenance operation(s) based on the identified data. In response to a determination that the current state does not satisfy one or more chamber maintenance criteria, a set of subsequent maintenance operations to be performed to cause the current state of the process chamber to satisfy the criteria is identified. Performance of the set of subsequent maintenance operations is initiated at the process chamber.

The invention relates to an audio-based slag splashing state monitoring method and system, and the method comprises the steps: collecting real-time data needed by model calculation in a slag splashing process after the starting of the slag splashing of a converter; analyzing and processing the collected real-time data; setting a slag rising threshold value, observing and learning sound intensity time sequence characteristics during slag rising, establishing a slag rising point automatic detection model, and determining a slag rising point; establishing a slag splashing early warning point detection model, and determining an early warning point; and establishing a slag splashing endpoint detection model, determining a slag splashing endpoint, and realizing automatic forecasting of the slag splashing endpoint. According to the method, the audio signal in the slag splashing process of the converter is monitored in real time, the audio time sequence characteristics are extracted, and the detection of the slag starting point and the accurate forecasting of the slag splashing end point are realized in combination with a machine learning algorithm and an end point prediction technology. According to the method, the production efficiency in the slag splashing process of the converter is improved, the manual intervention cost is reduced, the negative influence on the service life of the converter due to overlong slag splashing time is avoided, and a powerful technical support is provided for continuous progress of the steelmaking industry.

Information processing device that includes: an information acquisition unit that acquires reliability degradation factor state information in a chemical mechanical polishing process of a substrate performed by a substrate processing device, the reliability degradation factor state information including at least one of wear state information indicating a wear state of components of the substrate processing device and processing state information indicating a processing state during polishing; and a state prediction unit that predicts reliability information of a polishing endpoint detection function for the reliability degradation factor state information by inputting the reliability degradation factor state information acquired by the information acquisition unit into a learning model that has been trained through machine learning to learn a correlation between the reliability degradation factor state information and reliability information of the polishing endpoint detection function that indicates reliability of an endpoint detection function that detects that the chemical mechanical polishing process has reached an endpoint.

The present application provides a method, apparatus, device and storage medium for voice endpoint detection, wherein the voice endpoint detection method can determine whether the audio frame contained in the audio data to be detected is a silence frame, a noise frame or a voice frame, that is, the present application can detect the relatively accurate attributes of the audio frame contained in the audio data, and perform detection of the voice front-end point and the voice back-end point on this basis, so as to obtain relatively accurate detection results. On the basis of realizing voice endpoint detection, the present application can obtain the recognition text of the voice segment, and can determine the semantic scene of the recognition text based on the semantics of the recognition text, and then can set a suitable post-silence timeout threshold based on the semantic scene of the recognition text, thereby triggering a post-silence timeout event based on the suitable post-silence timeout threshold to improve the user experience.

The invention relates to the technical field of medical information processing, in particular to a medical voice instruction semantic analysis and form automatic filling system, which comprises the following modules: a voice recognition and analysis module, which is used for collecting a medical voice instruction, carrying out noise reduction processing and endpoint detection on a voice signal, and sending the voice signal to an automatic form filling module; generating a recognition text with confidence by using a voice recognition model customized by a medical scene, and performing named entity recognition and slot extraction on the recognition text to generate candidate semantic units; and the constraint generation module is used for extracting clinical logic constraints of the form fields according to the candidate semantic units. According to the method, constraint semantic representation is generated in combination with violation degree calculation, matching is completed under logic constraints by using a global optimization algorithm, and high-precision mapping of medical voice instructions and form fields is realized; not only is the integrity of semantic information ensured in the recognition stage, but also mistaken filling is effectively avoided through clinical logic constraints in the matching stage.

The application relates to a container endpoint detection method and device, a storage medium and equipment. A least square method is used to construct a straight line equation of n sampling points of laser scanning data of a to-be-detected container. Whether a sampling point after the nth sampling point belongs to the same straight line equation is judged to obtain an offset sampling point. The coordinate information of the first sampling point of the straight line equation and the coordinate information of the offset sampling point are used to obtain the coordinate information of a straight line endpoint. The straight line equation of the sampling point after the offset sampling point is reconstructed, and the coordinate information of the endpoint of the straight line segment corresponding to the straight line equation is obtained. Therefore, the position information of each endpoint of the to-be-detected container can be obtained according to the coordinate information of the endpoint, the user can quickly locate the endpoint position of the container according to the position information of each endpoint, and the detection efficiency of the container is improved.

The present application discloses an image processing method, system, electronic device and storage medium. Among them, the image processing method includes: inputting a first target magnetic resonance image into a brainstem generation model for modeling processing to obtain a three-dimensional brainstem image, and obtaining a target brainstem image according to the three-dimensional brainstem image; inputting the target brainstem image into a target boundary point detection model for detection processing to obtain boundary point coordinates; obtaining the midbrain area and pons area according to the boundary point coordinates, target point coordinates and target brainstem image; inputting a second target magnetic resonance image into an endpoint detection model to obtain endpoint coordinates, and then calculating the middle cerebellar peduncle data according to the endpoint coordinates; obtaining the superior cerebellar peduncle data according to the second magnetic resonance image, and finally obtaining target data according to the midbrain area, pons area, middle cerebellar peduncle data and superior cerebellar peduncle data. The image processing method of the embodiments of the present application can improve the measurement efficiency and measurement accuracy.

The semantic perception real-time voice endpoint detection method is applied to a voice interaction system, and comprises the following steps: receiving a voice stream of a user and a voice stream sent by the system from at least two audio channels; performing causal voice coding on the voice stream of the user and the voice stream sent by the system to obtain corresponding double-path voice representation; a causal two-way modeling architecture is constructed, and the causal two-way modeling architecture is realized by adopting an autoregressive prediction network and is used for receiving two-way voice representation and modeling a speech round conversion dynamic state between a user and a system through an embedded causal constraint mechanism and a cross-channel interaction mechanism; based on the output of the autoregressive prediction network, jointly predicting the category of a speech round conversion event and the remaining time from the current moment to the end of the speech round; and a smoothing processing mode is used to correct the predicted speech wheel conversion event type, and a final endpoint detection result is determined. According to the method, the endpoint detection stability in a complex dialogue scene can be improved.

The invention discloses an intelligent audio product with a call inter-translation function, and relates to the technical field of intelligent audio products. The intelligent audio product with the call inter-translation function comprises a wireless receiving and transmitting module, a translation module, a sound receiving module, a sound playing module and a communication module, the wireless receiving and transmitting module is connected with the sound receiving module, the sound receiving module is connected with the translation module, and the translation module completes translation based on voice signals based on voice information transmitted by the sound receiving module. According to the invention, content translation is completed through cooperation of each characteristic parameter with the Internet third-party translation server, local translation or cloud translation, information contained in the signals is concentrated while real-time translation is realized, the calculation amount in the voice signal processing process is greatly reduced, and the voice processing efficiency is improved. The problem that the translation difference of a traditional endpoint detection algorithm is large under the condition of a low signal-to-noise ratio is overcome, and the emotion color of a caller is still achieved while the translation timeliness is guaranteed.

This invention discloses a method for accurate audio warning identification based on hybrid features. This method analyzes collected audio speech to determine the issuance time of the warning signal, thereby accurately evaluating the timeliness of the audio warning. First, a double noise reduction method using logmmse-spectral subtraction is employed to filter out noise information in the recorded speech. Next, endpoint detection based on short-time energy is used to mark the effective speech segments in the test speech. Then, MFCC features and waveform polynomial features are extracted from each frame of the effective speech segment. Subsequently, the two features are used as inputs to two channels of a convolutional neural network, and the outputs of the two channels are summed to obtain the hybrid features. Finally, the hybrid features are used as input to a softmax function, and the speech segment containing the target speech (audio warning signal) is determined by the maximum probability value. The starting position of this speech segment is the issuance time of the warning signal.

To provide a method for manufacturing a polishing pad with an endpoint detection window (window member) that exhibits excellent adhesion between the back surface of the window member and the adhesive surface of the second sheet. [Solution] A polishing pad comprising a first sheet having a first through-hole, a second sheet having a second through-hole, and a window member, wherein the first sheet has a polishing surface for polishing an object to be polished, the second sheet is bonded to the surface of the first sheet opposite to the polishing surface, the outer periphery of the second through-hole is formed to cover the outer periphery of the first through-hole, and the window member is positioned inside the first through-hole and on the second sheet such that the uppermost part of the window member is lower than the polishing surface; and a method for manufacturing a polishing pad comprising the steps of: preparing the polishing pad, and placing a spacer of a predetermined thickness on a press platen under the polishing pad around the polishing pad, and pressing the polishing pad with a press machine.

The invention provides an end point detection method and equipment based on multiple light beams. The method is applied to semiconductor processing. The method comprises the steps that the wavelength # imgabs3 # of any incident light beam in multiple light beams is calculated through the target wafer film thickness # imgabs0 # corresponding to a wafer grinding end point, the refractive index # imgabs1 # of a wafer film and the incident angle # imgabs2 # of the light beams; the wavelength # imgabs4 # and the wavelength # imgabs5 # are at least utilized to establish a corresponding relation curve of the wafer film thickness and the reflectivity, the reflectivity changes periodically along with the thickness, and a point corresponding to the target wafer film thickness # imgabs7 # in the corresponding relation curve established by the wavelength # imgabs6 # is within a preset range in the curve; and irradiating at least two light beams containing the wavelength # imgabs8 # and the wavelength # imgabs9 # to the surface of the wafer film, and monitoring whether the thickness of the wafer film reaches the target wafer film thickness # imgabs11 # or not by utilizing the reflectivity corresponding to the target wafer film thickness # imgabs10 # and the corresponding relation curve. According to the invention, the end point detection sensitivity and the measurement accuracy are improved.

The utility model provides a light acquisition mechanism for an etching machine end point detection system, which relates to the field of etching machines and comprises a supporting seat, a convex lens assembly and an optical filter assembly. The supporting seat comprises a containing cylinder; the convex lens assembly and the optical filter assembly are detachably fixed in the accommodating cylinder; the optical filter assembly comprises a shell with a cylindrical through hole, an optical filter and a clamp spring; an annular baffle is arranged at the opening part of the cylindrical through hole; an annular groove is formed in the inner wall of the cylindrical through hole; the inner diameter of the ring of the snap spring is smaller than the diameter of the cylindrical through hole of the shell; the clamp spring is detachably fixed in the groove; the optical filter is arranged between the clamp spring and the circular ring baffle plate; and the two cylindrical bottom surfaces of the optical filter respectively abut against the snap springs and are fixed with the circular ring baffle plate. The light acquisition mechanism for the etching machine end point detection system provided by the utility model solves the problem that an optical filter is inconvenient to replace in the prior art, and the optical filter can be conveniently and stably replaced.

The application belongs to the field of mechanical processing, and discloses an endpoint detection method and system for polishing of high-hard-brittle materials, which comprises the following steps: step 1, constructing a mathematical model of chemical mechanical polishing; step 2, building a test platform to collect force and torque signals; step 3, designing a circuit of an edge computing platform; step 4, performing a parameter control experiment; step 5, developing a host computer system; step 6, finally obtaining a training set of a high-hard-brittle material processing state prediction model based on force and torque signals; step 7, training the processing state prediction model; and step 8, using the trained processing state prediction model to perform endpoint detection. The application uses polishing torque to perform online real-time detection on the polishing state and uses a state feedback method, so that the polishing processing efficiency can be improved, the polishing quality and the polishing yield can be improved, and the cost can be reduced.

The application provides a regulation voice segmentation method based on speech recognition and endpoint detection, which is applied to air traffic control voice audio stream segmentation, and comprises the following steps: step 1, constructing a punctuation model based on speech recognition and a speech endpoint detection model; step 2, using the punctuation model based on speech recognition to recognize the audio data stream of the regulation voice, and outputting the corresponding text and sentence end identifier of the audio data stream; step 3, using the speech endpoint detection model to judge the speech starting point and ending point contained in the audio data stream of the regulation voice; step 4, segmenting the audio data stream of the regulation voice into audio segments; and step 5, applying the audio segments as data materials to the speech recognition process of the air traffic control system. Through the combination of speech recognition and endpoint detection, the application realizes the automatic segmentation of the air traffic control voice audio stream, and improves the accuracy and efficiency of the segmentation.

An endpoint detection system for enhanced spectral data collection is provided. An optical bundle is coupled to a light source. The optical bundle includes an emitting optical fiber and a receiving optical fiber disposed at a pairing angle relative to the emitting optical fiber. The optical bundle is coupled to a collimator assembly that receives a light beam of incident light from the emitting optical fiber and directs spectral components of the light beam to first and second portions of a substrate surface. The collimator collects reflected spectral components produced by the spectral components directed to the substrate surface. The collimator assembly transmits the reflected spectral components to the receiving fiber, which transmits the reflected spectral components to a light detection component. A processing device coupled to the light detection component determines a reflectance of the substrate surface based on the reflected spectral components.

A technique for threat response associated with an endpoint detection and response (EDR) system. The system uses a combination of automated observable detection, threat intelligence enrichment, graph

The invention discloses a livestock respiratory rate estimation method, and relates to the technical field of animal health monitoring. The method comprises the following steps: collecting a livestock lung sound stream, segmenting into segments according to a fixed window length, and carrying out initial noise reduction; framing endpoints of the segments are detected, starting points and ending points of inspiration segments are marked, and after a sequence is obtained, a small subset and a large subset are divided according to the number of the inspiration segments; screening the effective inspiration section by using the expiration average duration time in the less set, and screening the effective inspiration section by using the difference value between the time interval of two adjacent inspiration sections and the expiration average duration time and the standard difference value in the more set; after the effective inspiration sections are screened out, the breathing frequency is calculated according to the ratio of the time interval between the head and tail inspiration sections of the target section to the number of effective breathing cycles. The accuracy of livestock respiratory rate detection can be effectively improved.

The application provides a speech endpoint detection method, device and equipment and a storage medium. The speech endpoint detection method comprises the following steps: obtaining target data, wherein the target data comprises target audio data; inputting the target data into a pre-trained audio frame classification model to obtain a classification result of an audio frame of the target audio data, wherein the audio frame classification model is trained by taking first audio data with noise in one or more categories of a plurality of noise categories as a training sample, taking a real category of the audio frame of the first audio data in a plurality of dimensions as a sample label, and the plurality of dimensions comprise a speech dimension and noise dimensions corresponding to the plurality of noise categories; and determining a speech endpoint according to the classification result of the audio frame of the target audio data. The speech endpoint detection method provided by the application can detect an accurate speech endpoint.

The purpose of the present application is to provide an endpoint detection method and device for audio, which obtains the target audio to be processed and its corresponding total duration; based on the total duration of the target audio, sequentially performs frequency domain transformation, feature extraction and effective feature filtering on the target audio to obtain the effective feature vector and sampling sequence number corresponding to each sampling moment in the target audio, and the absolute time difference between two adjacent sampling moments; based on the effective feature vector corresponding to each sampling moment, respectively predicts the category of each sampling moment and the relative deviation between each sampling moment and the moment of the true endpoint of the target audio; based on the absolute time difference and the sampling sequence number, category and relative deviation between the moment and the true endpoint of the target audio corresponding to each sampling moment, determines the moment of the starting endpoint and the moment of the ending endpoint of the target audio, thereby realizing accurate endpoint detection of audio with strong noise interference, with low computational complexity and high speed.

A technique for threat response associated with an endpoint detection and response (EDR) system. The system uses a combination of automatic observable detection, threat intelligence augmentation, graph analysis, and supervised machine learning to machine-predict analyst behavior in classifying EDR alerts (as "true" or "false" positives) and supports either (i) automatic suppression of alerts that the system classifies with sufficient confidence as either true or false, or (ii) providing recommendations to analysts for alerts that cannot be so classified to expedite their actions. The automatic detection of observables for graph-based feature detection, coupled with automatic processing of alerts when possible, significantly reduces the overall analyst workload for the EDR system. Furthermore, and even when the machine-based predictions are not sufficiently reliable to allow bypassing the analyst, the system provides additional context and augmentation to the analyst, facilitating faster (or at least more efficient) alert handling.