112 results about "False recognition" patented technology

The invention relates to a data correction method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring an original structured statement and original annotation information; determining knowledge data corresponding to the original structured statement in a target knowledge database based on a metadata query engine; the knowledge data comprises graph relation metadata and semantic vector metadata; according to the data correction large model, the graph relation metadata and the semantic vector metadata, error recognition is carried out on the original structured statement and the original annotation information, data correction is carried out on error data obtained through error recognition, and a target structured statement and target annotation information are obtained. By adopting the method, the accuracy of structured statement and annotation error correction can be improved.

A system and method for automated defective data record resolution is disclosed. The method is performed by a remote digital claim remediation service operating within a distributed network of computers. The method includes receiving adverse digital claim data comprising a plurality of data fields, where at least one data field contributes to an adverse action against a target digital claim. A feature extraction model processes the claim data to extract defect feature data. A machine learning model then computes defect inferences, identifying likely defect types and associated confidence values. The method further evaluates these inferences against predefined routing criteria, routing the claim data either to a remediation or disposal queue. For remediable claims, a machine learning model generates remediation inferences, which are applied to adapt and correct the claim. The defect-free claim is then resubmitted in real-time or near real-time, accelerating claim processing and enhancing system efficiency.

The invention provides an application packaging processing method, a storage medium and electronic equipment, and is applied to the technical field of software application. According to the method, a full-process closed-loop mechanism of the front stage, the middle stage and the rear stage of packaging is constructed; before packaging, application codes and key parameters are subjected to security detection; in the packaging process, grammar and semantic errors are deeply recognized through a regular engine, an abstract syntax tree and data flow analysis, and a diagnosis report is generated; and after packaging, executing an automatic test by utilizing the matched test case to obtain an application test result. Based on the diagnosis report and the test result, whether repackaging is needed or not is intelligently analyzed, packaging parameters and test cases are adjusted according to the error types and fed back to a safety detection link in a circulating mode, automatic error recognition and self-adaptive optimization are achieved, and therefore the safety and reliability of application packaging are effectively improved, and the development risk is reduced.

A method or system for enhancing vehicle identification accuracy. The system identifies a gap in a training dataset for training a license plate identification model. The gap represents underrepresented visual characteristics in misidentified license plates. A guidance prompt is generated based on the visual characteristics of the misidentified license plate. Condition embeddings are then generated from the guidance prompt, which are used to condition a diffusion model to create synthetic license plate images. The diffusion model is trained to receive a real license plate image, encode it into a vector, apply forward diffusion to add noise, and then apply reverse diffusion to remove the noise, resulting in a denoised vector that represents a synthetic license plate image conditioned by the condition embeddings. The denoised vector is then decoded to produce the synthetic license plate image. The synthetic images are then used to retrain the license plate identification model.

The present application relates to the technical field of face recognition, and particularly relates to a recognition method and system based on a dynamic face database, a computing device and a storage medium, comprising a face tracking server configured to acquire face information obtained by a channel camera and a regional camera, and to merge the face information obtained by the channel camera and the regional camera; a central face server configured to store face information left by a user when registering remotely, and to respectively identify face information collected by the face tracking server from the channel camera and the regional camera, and to send the identification results to a channel face server and a regional face server; the present application reduces the number of face databases by using face tracking and face recognition technology, forms a face server with multiple face databases, improves the recognition efficiency by accurately simplifying the face databases, and reduces the false recognition rate of face recognition.

The embodiment of the invention discloses a wrongly written character processing method, device and equipment, and the method comprises the steps: receiving text data of a target domain input by a user, the text data comprising special words of the target domain; performing word segmentation processing on the text data through a preset word segmentation strategy aiming at a target field, and performing identification processing aiming at special words of the target field on segmented words corresponding to the obtained text data to obtain target special words in the segmented words corresponding to the text data; carrying out wrongly written character recognition processing on the target proprietary word through a proprietary word error recognition model for the target field to obtain a first recognition result, and carrying out wrongly written character recognition processing on segmented words, except the target proprietary word, in the segmented words corresponding to the text data through a general word error recognition model to obtain a second recognition result; and based on the first recognition result and the second recognition result, determining wrongly written character information contained in the text data through fusion processing of the two different recognition results.

A user is assisted in performing a voice operation appropriately.A situation determination section determines a situation. A state control section controls a voice command appropriate for the determined situation to put the voice command into a receivable state. For example, the user is informed of what the voice command in a receivable state is, by means of display or voice output. The user can utter a voice command without performing a user action to prevent false recognition, such as the utterance of a wake word. This reduces the troublesomeness and burden of the user.

The application discloses a dynamic image feature capturing method based on deep learning, and relates to the technical field of image recognition. Through optimization of a deep convolutional neural network model, a convolutional layer in the deep convolutional neural network model which does not work or has a negative effect on a recognition result is removed, so that the calculation amount of the deep convolutional neural network model is reduced, and the accuracy of the deep convolutional neural network model is further improved on the basis of having completed training. Meanwhile, after a technical personnel deletes a feature marking frame of an error recognition through a tool, corresponding contrast video frames and original video frames are input as training samples of the deep convolutional neural network model, so that the recognition accuracy of the deep convolutional neural network model is more accurate.

This invention relates to the field of artificial intelligence education technology, specifically to a method and system for reviewing and providing feedback on physics, chemistry, and biology lab reports based on a multimodal large-scale model. The method includes: acquiring lab report question data, student answer data, and scoring standard data; vectorizing the question data and retrieving scoring criteria from a scoring knowledge base; selecting either a text-based or image-based scoring process based on the answer type, and calling a fine-tuned multimodal large-scale language model for scoring reasoning; calling an error identification module and a teaching suggestion generation module based on the scoring results; summarizing the complete review results and calculating the joint confidence level; and determining whether to push the results for manual review based on the confidence level. This invention can improve the accuracy, fine-grained understanding, and quality of teaching feedback in lab report review, enabling the self-evolution of the review system.

The method and system for detecting the humidity of needle-shaped tea based on image recognition disclosed in this application obtains the humidity detection results of the images of needle-shaped tea to be detected. When the number of images of the to-be-detected that fall within the range of the fixation processing value is greater than a first predetermined value, the corresponding humidity morphology is determined as the humidity morphology of the needle-shaped tea. Based on the morphology, clustering is performed on the humidity detection result library of the images of the to-be-detected needle-shaped tea to form a set of needle-shaped tea images corresponding to the humidity morphology. Data configuration filtering is performed to determine whether it belongs to a balanced or unbalanced configuration, and the fixation processing range value corresponding to the humidity morphology is determined. This solution only needs to obtain the current humidity detection result of the needle-shaped tea, and can directly determine the humidity morphology of the needle-shaped tea through the processing range value. It can be achieved without running a complex model, which is more efficient and accurate. Furthermore, it does not require the use of intermediate features such as color and brightness type, which reduces computational complexity and further reduces false recognition or missed recognition.

The present application relates to the cross technical field of speech signal processing, respiratory physiological monitoring and artificial intelligence, and specifically discloses a respiratory-speech pause signal analysis system and method. The present application synchronously collects speech and respiratory signals, extracts language-adapted pause parameters and physiological indexes, performs time sequence alignment and correlation analysis, and then fuses a prediction model and a true-false recognition model for intelligent analysis, thereby solving the problems that traditional heart-lung function detection relies on professional equipment and cannot be remotely and non-contactly monitored, and that existing speech analysis lacks a physiological coupling mechanism, leading to the inability to identify synthetic speech and insufficient cross-language adaptation, and realizing the dual ability improvement of non-contact respiratory physiological state evaluation and speech fraud recognition.

The present invention relates to a computer-implemented method to identify extremum locations of wave signals. The method comprises steps of obtaining a displacement series comprising displacement of each time frame of the wave signal, determining a total state series which maximizes a penalized total path as an optimized state series based on the displacement series, and determining positions of turning points of the wave signal based on the optimized state series. A transition penalty is introduced to the penalized total path to suppress false identification of noise-derived local extrema as turning points. A dynamic programming (DP) algorithm may be used to perform the computer-implemented method.

This invention discloses a method for constructing a virtual psychological supervisor based on multi-agent collaboration, belonging to the fields of digital mental health training and artificial intelligence technology. By constructing a multi-agent collaborative framework of "error-prone counselor agent - error-sensitive patient agent - error-correcting supervisor agent," and combining iterative optimization mechanisms guided by validators, high-quality "consultation dialogue - supervision feedback" aligned data pairs are generated. Based on data optimization of a large language model, a virtual psychological supervisor with error recognition, professional feedback generation, and skills assessment capabilities is ultimately constructed. This invention solves the problems of insufficient training data, inadequate feedback professionalism, and subjective assessment standards in traditional psychological counseling supervision, achieving safe, efficient, standardized, and scalable psychological counselor skills supervision, significantly improving supervision accuracy and training efficiency.

The application belongs to the field of circuit design, and particularly relates to a subgraph matching method based on an enhanced VF3 algorithm, a storage medium and equipment. The method first generates a type of graph structure data of a pattern circuit and type II graph structure data of the pattern circuit and a target circuit according to a netlist file, and establishes a mapping between nodes in different graph structure data. The type II graph structure data takes element ports as nodes, and takes electrical connection relationships between nodes, signal transmission directions and control logic as edges. Then, key nodes in the pattern circuit are identified and the priority of the remaining nodes is calculated. A priority queue is generated in which the key nodes are in front and sorted by priority. Finally, the priority queue is taken as a node search sequence, and a subgraph matching algorithm is used to match nodes in the pattern circuit and the target circuit in the type II graph structure data, so as to determine the pattern circuit contained in the target circuit. The application solves the problem of low search efficiency and error identification of the existing subgraph matching scheme.

The application provides a crowd statistical method and device based on confidence probability, comprising: obtaining an original picture for crowd quantity statistics; obtaining a feature map and a generated map corresponding to the original picture through encoding and decoding based on the original picture; inputting the original picture, the feature map and the generated map into a preset confidence probability model to obtain a confidence probability density map; inputting the feature map into a preset prediction neural network to obtain a prediction probability density map corresponding to the feature map; determining a real density map based on the confidence probability density map and the prediction probability density map, and performing crowd quantity statistics according to the real density map. In one aspect, the application reduces the data training model through encoding and decoding; on the other hand, the application screens the prediction probability density map by using the confidence probability density map, reduces the occurrence of false recognition, and thus improves the accuracy of crowd quantity statistics.

This application provides a training method for a fingerprint recognition network, a fingerprint recognition method, an apparatus, and a medium. The training method for the fingerprint recognition network includes: determining fingerprint feature matching maps based on a first initial fingerprint image and a second initial fingerprint image; forming a training sample set from multiple fingerprint feature matching maps; and training an original neural network based on the training sample set to obtain a fingerprint recognition network. The technical solution provided by this application solves the problem of difficulty in distinguishing whether two different fingerprint images belong to the same finger, thus reducing the false acceptance and false recognition rates of fingerprint recognition.

This application discloses a method, apparatus, device, and storage medium for determining head movements, relating to the field of head-mounted device technology. The method includes: converting initial angular velocity data collected by a gyroscope to a target coordinate system to obtain target angular velocity data; determining the target temporal features corresponding to the target angular velocity data, wherein the target temporal features include at least one of a first temporal feature corresponding to the first angular velocity data, a second temporal feature corresponding to the second angular velocity data, and a third temporal feature corresponding to the third angular velocity data; and matching the target temporal features with preset temporal features corresponding to each type of head movement to determine the target head movement type. Therefore, by extracting the temporal features of each of the three-axis angular velocity data and matching them with preset temporal features corresponding to various head movements, the detailed differences in head movements can be comprehensively captured, significantly improving the accuracy of head movement recognition and reducing the false recognition rate of similar movements.

The present invention provides a vehicle control device and a computer program capable of suppressing erroneous recognition as a pedestrian crosswalk when a road sign is detected. A vehicle control device is provided with a control unit that executes driving assistance control of a vehicle, and the control unit performs: analysis of imaging data obtained by imaging an environment around the vehicle; identifying a road present in the environment in a direction in which the vehicle travels; calculating a first distance in the width direction of the road; identifying a road surface identifier existing on the road; calculating a second distance in the width direction of a predetermined region including the road surface mark when it is recognized that the road surface mark is the predetermined pattern representing the pedestrian crosswalk; calculating a ratio of the second distance to the first distance; when the ratio is equal to or greater than a predetermined value, the driving assistance control unit recognizes a road surface mark as a pedestrian crosswalk and executes driving assistance control. And when the ratio is less than the predetermined value, recognizing the road surface mark as being not the pedestrian crosswalk, and not executing the driving assistance control.

The application discloses a fingerprint template updating method and device, an embedded device and a storage medium, relates to the field of fingerprint identification, and is used for an embedded fingerprint identification device. The method comprises the following steps: collecting a to-be-identified fingerprint image, extracting to-be-identified fingerprint features of the image, comparing the to-be-identified fingerprint features with a template library, and determining a target fingerprint template of a target finger; comparing the to-be-identified fingerprint features with identified fingerprint features in the target fingerprint template, determining to-be-replaced features of the target fingerprint template according to a feature comparison score and a feature template coverage area; calculating a replacement score of the target fingerprint template based on the to-be-identified fingerprint features and the to-be-replaced features, and performing template updating when the replacement score exceeds a score threshold. In the fingerprint identification process, the corresponding replacement score is determined in a manner of fingerprint feature matching and feature area comparison, the template is updated according to the score size, the fingerprint identification precision is improved, and the false recognition rate is reduced.

This invention discloses an encrypted file box that can be remotely unlocked via facial recognition. The encrypted file box is equipped with a recognition system. After recognizing the correct signal, the recognition system sends a command to a control unit embedded in it, which then controls the unlocking of the encrypted file box. Both the recognition system and the mobile client are connected to a network, and data is transmitted between them through account interaction. This invention uses facial recognition technology as the unlocking method, effectively improving the security level of the file box. Users can remotely perform facial recognition via a mobile client, enabling secure unlocking of the file box even when not physically present, providing great convenience. Furthermore, this invention enhances the accuracy of recognition and reduces the possibility of false recognition by extracting spatiotemporal features and biometric signals and combining these two types of information for feature fusion.

The application provides a kind of earphone knock detection method, the method comprises the following steps: obtaining the three-axis acceleration data of earphone of current acquisition window, and the change of acceleration is calculated based on three-axis acceleration data to determine dynamic knock threshold;If current instantaneous knock intensity is greater than dynamic knock threshold, it is determined that pre-knock event;If there is second pre-knock of instantaneous knock intensity greater than dynamic knock threshold in the first detection interval for continuous hit waiting after determining pre-knock event, it is determined that continuous hit event and records the number of hits. By calculating the change of acceleration based on the three-axis acceleration data to determine the dynamic knock threshold, the environmental adaptability is improved compared to the setting of fixed threshold, and the response intensity and time of the knock and the continuous hit time after the knock are filtered, effectively filtering short-time random vibration, such as false recognition caused by wearing adjustment or movement, improving user experience.

Arrangements for quantum enabled error identification are provided. A computing platform may train an artificial intelligence (AI) engine. The computing platform may extract metadata from one or more interconnected systems. The computing platform may use the metadata to map connections within the one or more interconnected systems. The computing platform may generate one or more dynamic markers. The computing platform may convert the metadata into qubits. The computing platform may generate one or more quantum enabled differential privacy components. The computing platform may inject the components into the one or more interconnected systems. A dynamic marker may be highlighted. An error may be identified based on the dynamic marker that was highlighted. An action may be identified based on the error. The action may be automatically executed by the computing platform.

A method (400) for using anti-context examples for personalizing a speech recognition model (132) includes receiving audio data (104) corresponding to an utterance (102) spoken by a user (10), and processing, using the speech recognition model, the audio data to generate a transcription (106) of the utterance. The transcription including a misrecognized phrase (144) that was misrecognized in the transcription by the speech recognition model. The method also includes receiving user-corrected text (141) including a corrected phrase (146) that replaces the misrecognized phrase that was misrecognized in the transcription. Based on the misrecognized phrase, the method includes generating an anti-context example (305) including anti-context text (310) containing the misrecognized phrase paired with text-to-speech (TTS) audio data (315) corresponding to a synthesized speech representation of the anti-context text. The method also includes personalizing the speech recognition model based on the anti-context example.

The invention relates to a method for rapidly extracting palm vein information from a palm print image, and belongs to the technical field of vein recognition and information security, and the method comprises the following steps: collecting the palm print image through a mobile terminal, converting the palm print image into an RGB data format, and then carrying out adaptive adjustment; carrying out filtering processing on the RGB three-channel image of the palmprint image after adaptive adjustment; extracting palmprint features, and performing feature fusion by using the palmprint features of the palmprint image in the R, G and B channels to obtain palmprint fusion features in the palmprint image; the palm vein information is extracted based on the palm print fusion features in the palm print image, the palm vein technology can be directly applied to the mobile terminal equipment, the similarity between the palm vein image and the palm vein image collected by a professional palm vein image collector is larger than 99%, the palm vein information is quite accurate, the accuracy of identity recognition is improved, and the user experience is improved. And the false identification rate is reduced.

The application relates to the technical field of BIM intelligent design and deep learning, in particular to a BIM model graph structure representation and model error identification method and device, wherein the method comprises the following steps: extracting semantic features, spatial features and topological features among various components according to component information of different types of components in a BIM model; generating corresponding graph structure representations based on the semantic features, the spatial features and the topological features, and constructing a training data set by using the graph structure representations; training a pre-trained graph neural network model by using the training data set to obtain a trained training model, and performing transfer learning by using the trained training model to identify an error identification target of the BIM model by using a transfer learning model. Thus, the problems that in the related art, the complex semantic, spatial and topological features of the BIM model bring difficult-to-ignore challenges, and the pre-training and transfer learning model cannot be perfectly suitable for the BIM model features are solved.

The application provides an anti-interference acoustic signal data extraction method based on energy data sensing, and belongs to the technical field of underwater acoustic signal processing. The method comprises the following steps: A / D sampling and digital quadrature demodulation are performed on a received acoustic signal to obtain a baseband signal; low-pass filtering and down-sampling are performed on the baseband signal to obtain a pretreated signal; the pretreated signal is matched filtered by using a reference signal; a key improvement lies in that a local energy estimation function of the pretreated signal is calculated, and the matched filtering result is normalized with the energy estimation function to obtain a normalized detection signal. Through energy sensing normalization, the application effectively suppresses the shielding effect of strong interference signals on weak target signals due to non-quadrature reference signals and the "near-end strong and far-end weak" scene, significantly reduces the false recognition probability, and improves the accuracy and reliability of extracting a long-distance and low signal-to-noise ratio target beacon signal in a complex underwater acoustic environment. The application is suitable for underwater navigation, positioning and communication systems.

This application relates to the field of satellite navigation and positioning technology, and provides a method, apparatus, receiver, and storage medium for ambiguity fixing. First, the ambiguity candidate subset is fixed using the LAMBDA algorithm to obtain the ambiguity fixing result. If the ambiguity fixing result meets preset conditions, the ambiguity fixing is determined to be correct, and the final ambiguity fixing solution is output. If the ambiguity fixing result does not meet the preset conditions, the ambiguity fixing is determined to be incorrect. Incorrectly fixed observations are identified and removed from the ambiguity candidate subset. Then, the LAMBDA algorithm is used again to fix the ambiguity candidate subset until the ambiguity is correctly fixed. This improves the accuracy of ambiguity fixing to meet the requirements of high-precision GNSS positioning.

This invention discloses a face recognition method, apparatus, device, and storage medium based on multi-view acquisition, relating to the field of image processing technology. The method includes: in response to detecting a target object in the face recognition area of ​​a tri-lens encryption device, acquiring a set of multi-view image frames captured by three cameras arranged in an arc for the target object's face; preprocessing each frame in the multi-view image frame set to extract facial feature information from each viewpoint; fusing the facial feature information from each viewpoint to obtain target feature information; and triggering the face recognition process of the tri-lens encryption device based on the target feature information. By implementing the technical solution of this invention, the dimensions of facial features are expanded, misidentification of similar faces at specific angles is avoided, the accuracy of face recognition is improved, and the false recognition rate and rejection rate are significantly reduced.

The invention provides a communication controller error detection method and system based on algorithm optimization, and is applied to the technical field of communication controllers, and the method comprises the steps: obtaining the real-time operation state parameters, including internal state parameters and external environment parameters, of a communication controller, and carrying out the correlation analysis based on the parameters, thereby achieving the error detection of the communication controller. The method comprises the steps of determining a risk assessment value of a target control data packet, then dynamically adjusting an error detection strategy according to the risk assessment value, including adjustment of a detection sensitivity threshold, switching of verification algorithm complexity or starting and stopping of an error correction function, entering a strategy locking state when the risk assessment value fluctuates violently, calculating an average value and adjusting the strategy at one time. Therefore, the method has the advantages that the detection strategy can be adaptively adjusted according to the real-time communication condition, and the verification parameters are optimized, so that the error recognition efficiency is improved.

This invention provides a method, server, and storage medium for anti-counterfeiting identification of corks. First, a cork image is acquired. Then, character codes are extracted from the cork image, and the original texture image of the cork is retrieved from a database based on these character codes. Next, the cork outline in the cork image is identified to determine the compression state of the cork. Finally, the anti-counterfeiting identification result of the cork is determined based on the cork image, the original texture image, and the compression state. By matching the cork image taken by the user with the original texture image stored in the database, and combining this with the compression state of the cork after it has been opened, the matching process is completed, avoiding misidentification caused by cork deformation and effectively improving the accuracy of anti-counterfeiting identification of corks.