Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

16results about How to "Recognizable" patented technology

Open sash unlocking identification method and system based on deep neural network and background modeling

The invention provides an open sash unlocking identification method and system based on a deep neural network and background modeling, and the system comprises a camera which is used for capturing a video stream of a door and window region in real time; the deep neural network module is used for automatically identifying the position and contour of a window and outputting bounding box coordinates and classification labels of a window area; the background modeling and motion detection module is used for detecting motion states of doors and windows; the geometric analysis module is used for calculating the intersection-to-union ratio and the Euclidean distance between the window boundary frame and the motion area so as to evaluate and verify the detection result; the locking state judgment module is used for judging whether the door and window are not locked or not according to the detection result of the continuous frames; and the alarm and feedback module is used for triggering alarm and sending specific position and state information to management personnel. According to the invention, by combining the deep neural network and the background modeling technology, automatic detection of the unlocked state of the large-scale glass curtain wall building door and window is realized, the traditional manual detection is replaced, and the detection efficiency and accuracy are greatly improved.
Owner:DECORATION CO LTD OF CHINA CONSTR 3RD ENG BUREAU +1

A WiFi cross-domain gesture recognition method based on domain-invariant feature extraction

PendingCN122657961ARealize cross-domain identificationsimple processChannel state informationA domain
The application provides a WiFi cross-domain gesture recognition method based on domain-invariant feature extraction, and relates to the technical field of WiFi cross-domain gesture recognition.The method comprises the following steps: acquiring channel state information data collected by a WiFi card; performing data preprocessing on the channel state information data to obtain a preprocessed image; inputting the preprocessed image into a domain-invariant feature extraction network fused with a channel space mixed attention module and a ConvNeXt network to extract domain-invariant features; and outputting a gesture recognition result through a classifier from the domain-invariant features.The method can effectively suppress domain shift interference without relying on target domain samples and without complex signal processing, thereby improving the accuracy and robustness of WiFi cross-domain gesture recognition.
Owner:UNIV OF ELECTRONICS SCI & TECH OF CHINA +1

A symmetrical cross-wing layout unmanned aerial vehicle system

This invention proposes a symmetrically arranged cross-wing unmanned aerial vehicle (UAV) system, relating to the field of UAVs. It includes a fuselage structure, a power module, and a flight control module. The fuselage structure comprises a cylindrical fuselage and four wings symmetrically arranged in a cross pattern around the cylindrical surface of the fuselage; the four wings are arranged in an X-shape or a cross shape. The power module includes motors and propellers. The motors are located at the wingtips, with their output shafts aligned with the fuselage axis and facing the tail of the fuselage. The propellers are connected to the motor output shafts. The flight control module is located inside the fuselage and is used to control the output speed of the four motors. This system combines the advantages of a quadcopter FPV and a fixed-wing UAV, generating lift from the wings while also cruising and sprinting in fixed-wing form. Furthermore, by changing the motor speeds, it can control the pitch, yaw, and roll attitude of the UAV, eliminating the need for separate control surfaces, thus simplifying the structure and reducing costs.
Owner:BEIJING HENGJU HONGTU TECHNOLOGY LLP (LLP)

An engineering cost auditing method and system based on multi-source data fusion

The application discloses a kind of engineering cost auditing method and system based on multi-source data fusion.Belongs to the field of engineering cost auditing, the present application obtains the engineering quantity list of the engineering to be audited, construction site space-time image and construction log text data, input model trained by multi-modal data, generates cost declaration and physical live characteristic vector, then calculates correlation score and extracts empirical segment features, and with cost declaration features fusion generates semantic fusion feature matrix;Finally, identify sub-item engineering category, calculate its Mahalanobis distance with process feature prototype vector, if exceed threshold value, then determine abnormal, generate the audit result containing abnormal cost item and construction period lack of empirical information.The present application effectively solves the problem that existing technology cannot accurately and efficiently audit engineering cost.
Owner:GUANGDONG OCEAN UNIVERSITY

PCVCs end-to-end speech recognition method based on ResNet-GAU model

The present invention provides a PCVCs end-to-end speech recognition method based on the ResNet‑GAU model, which belongs to the field of speech recognition technology, including: using ResNet to extract the time-frequency domain features of the speech signal to be recognized; through multiple series-connected GAU modules, the word order information of the time-frequency domain features is captured; the word order information is passed into the Denselayer, and the final classification prediction probability is obtained through the softmax layer after linear transformation in the feature space. Among them, the ResNet in the present invention uses the translation invariance and local correlation of CNN to extract the time-frequency domain information of the speech signal; GAU uses a gated single-head attention mechanism to better capture the long-distance dependency of the sequence to obtain a larger receptive field and context information, and also has a faster training convergence speed; CTC uses the introduction of blank and the generation of repeated tokens to solve the problem of hard alignment of speech signals and text labels.
Owner:CIVIL AVIATION FLIGHT UNIV OF CHINA

Multi-dimensional traffic grayscale routing method and system based on traffic white list

PendingCN121770835ARecognizableAchieved controllabilitySecuring communicationRouting decisionEngineering
The invention provides a multi-dimensional traffic grayscale routing method and system based on a traffic white list, and belongs to the technical field of data processing, and the method comprises the steps: receiving client request data containing a user identifier, channel information and at least one custom scene attribute; performing legality verification on the user identifier, the channel information and the custom scene attribute; matching the request data passing the verification with a preset multi-dimensional white list rule, the multi-dimensional white list rule at least comprising a user white list, a channel white list and at least one extension type white list, and the white lists of different dimensions have priorities; determining a target service version according to the matching result and the priority; and executing a routing decision based on the determined target service version, and forwarding the request data to the corresponding server, thereby realizing accurate identification and controllable distribution of the gray traffic.
Owner:CHANGSHA WEIFUTONG TECH SERVICE CO LTD

Semantic learning-based time series signal processing method, device and equipment

The present application relates to a method, device and equipment for processing time series signals based on semantic learning. The method includes: using a time series semantic generation module to sample, divide and weight the time series sequence; the feature extraction module uses a self-attention mechanism to capture the local correlation and global correlation between the weighted divided samples; using a contrastive learning module to calculate the contrastive learning loss of the encoded features, so that the model learns the common features between positive samples and the separable features between negative samples to obtain time series semantic features; the time series semantic features are input into the classification head of the preset time series target classification task to obtain the target recognition result. This method can realize the semantic feature learning of the time series sequence, improve the model's understanding of the time series semantic information contained in the time series sequence, and finally enable the model to obtain time series features with good representation, effectively improve the efficiency and accuracy of classification and recognition, and achieve good recognition performance under various interference conditions.
Owner:NAT UNIV OF DEFENSE TECH

Dynamic dense connection spatial-temporal feature decoupling network for recognizing cross-view gait

The invention discloses a dynamic dense connection spatio-temporal feature decoupling network for identifying cross-view gaits, which relates to the field of computer vision and comprises an initial feature processing module, a dynamic dense spatio-temporal decoupling feature extraction module and a feature enhancement processing module. Sharing of shallow network features and deep network features is realized by using dense spatial-temporal feature decoupling blocks and splicing operation, so that the problem of insufficient characterization capability is solved; meanwhile, a convolutional block attention enhancing mechanism is used for enabling the network to pay attention to more important gait features; and finally, the five-dimensional features are mapped and processed into a plurality of transverse features in a feature enhancement processing module, and batch standardization is performed on the features to enhance the representation capability and the model generalization capability, so that the correlation between shallow and deep features can be mined, and the problem of insufficient information expression capability is relieved.
Owner:HEFEI UNIV

Radar puzzle data abnormal echo identification method and system based on semantic segmentation

The invention provides a radar jigsaw data abnormal echo recognition method and system based on semantic segmentation, and relates to the technical field of radar jigsaw data quality control, and the method comprises the following steps: constructing a radar jigsaw image data set based on a radar combined reflectivity image; marking abnormal echo images in the radar jigsaw image data set, and constructing a radar jigsaw abnormal echo training data set; introducing a middle-level feature extraction layer and a simple attention module into the semantic segmentation network to obtain an improved semantic segmentation network; based on the improved semantic segmentation network, constructing a radar puzzle data abnormal echo recognition model; and training a radar puzzle data abnormal echo recognition model based on the radar puzzle abnormal echo training data set, and performing abnormal echo recognition on the to-be-recognized radar puzzle data to obtain an abnormal echo recognition result. According to the technical scheme, technical support is provided for further improving the data quality and the application level of the radar jigsaw data.
Owner:CHINA METEOROLOGICAL ADMINISTRATION WUHAN RAINSTORM RES INST

Deep learning based multi-modal fusion wireless signal automatic modulation recognition method

This invention discloses an automatic modulation identification method for multimodal fusion wireless signals based on deep learning, belonging to the field of communication technology. First, a publicly available dataset is selected as the channel input data and merged according to modulation type to form a set of time-domain I / Q data. Then, a constellation diagram is created for each set of time-domain I / Q data, and the data and corresponding constellation diagrams are segmented according to the signal-to-noise ratio to form a training set. Next, a multimodal fusion neural network model based on the time-domain I / Q data and constellation diagram data is constructed and input into the training set for iteration, adjusting the learning rate until the network model reaches stability. Finally, a newly acquired radio signal to be identified is input into the trained multimodal fusion neural network model, which automatically outputs the modulation identification type of the wireless signal. This invention can more fully extract the fusion features of the signal in the time domain and the corresponding constellation diagram to optimize the information loss caused by downsampling, thereby obtaining more accurate signal phase and amplitude information.
Owner:BEIJING FORESTRY UNIVERSITY

Bridge damping ratio rapid identification method based on stationary detection vehicle

This invention discloses a rapid bridge damping ratio identification method based on a stationary inspection vehicle, belonging to the field of bridge structure vibration testing technology. The method includes: parking the inspection vehicle at the mid-span of the bridge under inspection; exciting the bridge vibration using a passing vehicle and then driving away; collecting the acceleration response of the inspection vehicle in the attenuation segment; establishing a dynamic model of the attenuation segment of the vehicle-bridge coupled system; constructing a two-degree-of-freedom joint state model, internalizing the vehicle-bridge interaction force as a linear combination of system state variables; extracting the effective attenuation segment through signal preprocessing and introducing a dual anti-noise mechanism of adaptive signal-to-noise ratio weights and Huber weights; and employing a profiled two-stage parameter identification strategy, first locking the frequency and modal mass, and then finely scanning and identifying the bridge's first-order damping ratio. This invention requires only a single inspection vehicle and an acceleration sensor, eliminating the need to deploy sensors on the bridge structure. It has strong anti-interference capabilities, high identification accuracy, and is suitable for rapid inspection of small-to-medium span bridges.
Owner:CHANGSHA UNIVERSITY OF SCIENCE AND TECHNOLOGY

Dual antibody sandwich elisa detection kit based on nanobody-hrp fusion protein of porcine delta coronavirus and its detection and application

PendingCN122506158ARecognizableFunctional
The application belongs to the technical field of biological immune detection, and discloses a pig delta coronavirus double-antibody sandwich ELISA detection kit based on a nano antibody-HRP fusion protein as well as detection and application thereof. A PDCoV nucleocapsid protein (N protein) is taken as a detection target, and two strains of nano antibodies capable of recognizing different spatial epitopes of the PDCoV N protein are taken as core recognition elements to construct a sandwich type immune detection system. One strain of the nano antibodies is taken as a solid-phase capture antibody, the capture antibody is an unmodified nano antibody, is fixed on the surface of a solid-phase carrier, and is used for specifically capturing the PDCoV antigen in a sample. The other strain of the nano antibodies is gene-fused with a horseradish peroxidase to form a detection antibody with antigen recognition and signal catalysis functions. The minimum detection limit reaches 2.75 x 10 2 TCID 50 / 100 muL. The virus shedding can be detected 24 hours after the piglet is infected.
Owner:SICHUAN UNIV

Method and device for obtaining handwriting information

The present application discloses a method and device for acquiring handwriting information, the method comprising: acquiring multiple groups of collected data collected from the pen at multiple collection moments during the writing process of the pen, wherein each group of collected data includes the motion-related data of the pen and the initial image obtained by photographing the pen tip; using the motion-related data, determining the position representation information of the pen tip at each collection moment; based on the position representation information at each collection moment, transforming the initial image at each collection moment to the same space to obtain the transformed image at each collection moment; wherein the transformed image at each collection moment is used to identify the writing handwriting of the pen. Through the above method, the present application can be written on ordinary paper without relying on a matching handwriting book, thereby realizing complete handwriting recognition and recording.
Owner:IFLYTEK CO LTD

An endoscopic polyp detection method based on improved YOLOv8

This invention discloses an endoscopic polyp detection method based on an improved YOLO v8. Addressing the difficulty in detecting polyps in endoscopic images and the tendency to miss them during endoscope movement, the YOLO v8 detection model is improved in three ways: Channel attention (CA) is added after several C2f modules in the Backbone and Head layers of the original model to strengthen the focus on key features and improve the model's representation ability; a small target detection layer is added to the feature pyramid output layer to improve small target detection capability through position regression of small-sized feature maps; and WloU is used instead of CloU in the original model, which can adaptively adjust the loss term according to the target size and shape, improving the accuracy of small target localization. Based on this, an endoscopic polyp detection method is proposed. The key steps of the method are as follows: First, the relevant public dataset is obtained, preprocessed, and appropriately divided into training, testing, and validation sets; second, a detection model is constructed using the improved YOLO v8; finally, the model parameters are initialized for model training, and the detection accuracy and generalization ability of the model are evaluated using the test set. This method can accurately identify polyp areas in endoscopic images, reduce the false negative rate, and improve the accuracy and reliability of endoscopic examinations, thereby assisting medical staff in improving the accuracy of pathological analysis and disease diagnosis, while also increasing the work efficiency of medical staff.
Owner:HOHAI UNIV

Automobile decoration front face

An automobile decoration front face is characterized in that light spot units which transmit light or reflect light to form light spot areas are arranged on a light spot plate, the light spot plate is attached to a background plate, hollowed-out units are arranged on the background plate, the hollowed-out units correspond to the light spot units in position and are flush with the light spot units in outline, and chromatic aberration exists between light spots formed by the background plate and the light spot units. The background plate and the light spot plate are matched with each other, the hollowed-out units correspond to the light spot units in position and are flush with the light spot units in outline, the background plate shields the areas, except the light spot units, on the light spot plate, the light-transmitting or light-reflecting light spot units with clear boundaries can be formed on the light spot plate, and therefore the aesthetic design effect of the automobile front face is improved.
Owner:HEFEI CONVER HLDG

Image labeling grayscale compensation method, system, electronic device and medium

PendingCN122597244ARecognizableEliminate grayscale drift
The application discloses a kind of image mark gray compensation method, system, electronic equipment and medium, the method relates to computer vision detection field, the method specifically includes: obtaining the image mark containing image to be compensated image;Based on the grid side length of pre-set, the preset region containing image mark in to-be-compensated image is carried out grid down sampling processing, and down sampling image is obtained;Wherein, grid side length is less than the smaller value in width, height of image mark in image plane;The gray difference value of the gray value of each down sampling pixel and pre-set target gray value is obtained, and all gray difference values are mapped and arranged according to the spatial coordinates of corresponding down sampling pixel, and difference value image is generated;Difference value image is carried out spatial interpolation up sampling processing, and compensation amount image is obtained;Based on compensation amount image, pre-set region is compensated in gray scale.The method provided by the application image mark still has discernibility after being compensated in gray scale.
Owner:ZHUHAI BOJAY ELECTRONICS