19results about How to "Improve feature extraction accuracy" patented technology

The invention provides an access control system based on face recognition and a recognition method, and the system comprises a multi-modal image collection subsystem which integrates a visible light camera, a depth camera and an infrared thermal imaging camera and is used for synchronously collecting a visible light image, a depth image and an infrared thermal imaging image of a face; the image preprocessing unit is used for carrying out corresponding preprocessing on the image; the feature extraction network module is used for extracting image features; the multi-modal feature fusion and alignment module is used for calculating a mutual information matrix by adopting a feature fusion strategy based on mutual information maximization and adaptive weighting, determining an adaptive weight and performing weighted summation, and performing feature alignment by applying a method based on canonical correlation analysis; and the face recognition classification and verification module verifies the image by adopting a multi-level fusion strategy. According to the method, the false recognition rate and the rejection rate are effectively reduced, the strong anti-cheating capability is achieved, and cheating attacks of various fake face means can be accurately recognized and resisted.

The invention relates to the technical field of touch screens, and discloses a multi-touch mode switching method, device and equipment for a touch screen, and the method comprises the steps: carrying out the minimum error processing of a touch signal collected by a touch screen, and obtaining a target feature component; inputting the target feature component into a multi-layer touch mode classifier for feature mapping to obtain a current touch mode feature; performing attenuation characteristic analysis and adaptive threshold adjustment on the target characteristic component to obtain a state transition parameter; carrying out delay response modulation and regularization zero interference on the response time, the response intensity and the response area of the touch point to obtain screen response change characteristics; performing fuzzy logic decision of a continuous sampling period to obtain a touch mode switching feasibility judgment result; the closed-loop control is performed on the response parameter of the touch screen to obtain the mode switching delay compensation control quantity, so that the reliability of touch mode switching is improved, the false triggering phenomenon is avoided, and the response delay in the mode switching process is remarkably reduced.

The application is suitable for the medical technology field, and particularly relates to a speech recognition method and device based on artificial intelligence, computer equipment and medium. The application obtains a first speech enhancement matrix and a second speech enhancement matrix missing different semantic information by randomly shielding rows and columns of a mel spectrum matrix; calculates a metric sub-loss to perform self-supervised training on an encoder according to the first speech frame features and the second speech frame features extracted by the encoder; obtains speech fusion features and inputs the speech fusion features into a decoder to obtain mapped characters, combines preset characters to calculate a prediction loss to perform supervised training on a speech recognition model, weights and adds the prediction loss and the metric loss according to the number of zero characters and non-zero characters to obtain a target loss to train the encoder and the decoder, and combines the self-supervised training mode and the supervised training mode to improve the recognition accuracy of the speech recognition model, greatly improving the instantaneity, convenience and accuracy of information input in the medical technology field.

The application discloses a small target detection method and system for unmanned aerial vehicle aerial image, relates to the technical field of computer vision and unmanned aerial vehicle target detection, and generates a C2f_iAFF module by embedding an iterative attention feature fusion mechanism iAFF in a C2F convolution module of a YOLOv8 network, trains the C2f_iAFF module by using an STDIoU loss function which is optimized based on MPDIoU and is suitable for small target detection characteristics, adds a small target detection head to an original network structure of the YOLOv8 to generate an improved network structure, and finally obtains a small target detection model which is used for detecting unmanned aerial vehicle aerial images, obtaining a small target detection type and result, and improving the precision and speed of small target detection.

The invention belongs to the technical field of high-voltage lightning stroke fault prevention, and more specifically relates to a lightweight lightning image instance segmentation method. The method comprises the steps that firstly, lightning images existing in a target area in thunderstorm days are obtained and cleaned, images which do not meet conditions are removed, and the reserved images should be exposed moderately; then, inputting the cleaned lightning image data into a YOLO-lightning network model, and carrying out instance segmentation on the lightning image data; and finally, outputting contour information of the to-be-identified lightning in the image. According to the method, the problems of insufficient precision and limited feature extraction capability in complex form target instance segmentation in the prior art are solved.

The present invention discloses a point cloud classification and segmentation method based on fast sampling and attention mechanism, comprising the steps of: loading a point cloud classification data set D c And point cloud segmentation dataset D s ; Construct a point cloud classification and segmentation network based on the PointNet++ network, and construct a downsampling layer to replace its original sampling layer, and construct a segmentation network based on the attention mechanism N St Replace the original segmentation network, in N St Improve the interpolation operation in , and add a self-attention encoder in the skip connection; use D c and D s The point cloud classification and segmentation network is trained to obtain a point cloud classification and segmentation model, which is used in the classification or segmentation tasks of point cloud data. The present invention can reduce the complexity of the downsampling process, focus on and extract the space and related information in the point cloud, and overcome the problem of inaccurate interpolation caused by the selected neighborhood size and point distribution during interpolation. Therefore, the present invention can improve sampling efficiency, feature extraction accuracy and segmentation accuracy.

The invention discloses an artificial intelligence-driven burn wound assessment method and system, and relates to the technical field of medical information processing and intelligent diagnosis, and the method comprises the following steps: S001, employing a fixed imaging attitude to collect continuous short-sequence wound image frames, calculating a brightness change curve of adjacent image frames, generating a light and shade rhythm track, and drawing a beat frequency prompt band, the method is used for marking a light source flicker rule and establishing an initial reference baseline of brightness change. Through brightness breakpoint identification, error shooting parameter calculation, light source phase jitter and inverse rhythm optical shutter regulation and control, active separation and reverse coupling of illumination rhythm and sampling rhythm are realized, periodic flicker interference is effectively eliminated, the brightness stability and texture reduction degree of a wound image are improved, an artificial intelligence evaluation result is more accurate and reliable, and the artificial intelligence evaluation efficiency is improved. And the objectivity and safety of clinical diagnosis are improved.

The invention provides an abnormal capital collection account group identification, clustering and label labeling method based on capital flow data, and belongs to the technical field of financial risk prevention and control and data mining. A time sequence LSTM-structured self-encoding fusion AI feature extraction module is customized; the method comprises the following steps: capturing time sequence features such as periodic transfer and large-amount concentrated transfer through an LSTM attention layer, extracting structured features such as cross-regional association through an auto-encoder with a risk penalty term, and performing weighted fusion to obtain 12-dimensional AI features; and then semi-supervised K-means is used to identify a suspicious account group, spectral clustering is used to divide a case cluster, a random forest is used to label a'capitator / investor 'label, and the precision is ensured through three-layer verification. The method solves the problems of incomplete artificial feature coverage and poor universal AI adaptability in the prior art, and is suitable for abnormal capital investigation of financial supervision departments.

The present application relates to the field of industrial robots and machine vision technology, specifically a multimodal visual target recognition method for a six-axis industrial robot, comprising: synchronously acquiring image data, depth data and kinematic state data; generating ideal reference images and ideal reference depths based on a target model library, imaging parameters and candidate poses; injecting occlusion, shadow and reflection interference based on kinematic states, light source parameters and reflection model parameters to generate theoretical disturbed simulation data; extracting real residual errors and theoretical residual errors, and calculating a normalized cross-correlation coefficient in a local window, and performing masking rejection on noise areas; inputting the rejected residual errors into a recognition network to output target categories and corresponding target poses; and dynamically calibrating light source parameters, reflection model parameters and correlation thresholds according to re-projection deviations; the present application can improve recognition stability and pose output reliability in complex metal workpiece stations.

The application provides a construction worker unsafe behavior intelligent identification method and early warning system, the method comprises the following steps: obtaining a to-be-detected image of a construction site; inputting the to-be-detected image into a construction worker unsafe behavior identification model to obtain an unsafe behavior identification result of a construction worker; wherein the unsafe behavior at least includes an unworn state, a wrong wearing state or a blocked state of personal protective equipment; and the construction worker unsafe behavior identification model is a neural network model for target detection. The detection accuracy is significantly improved, and the fine-grained identification capability is enhanced. Since a learnable position coding component is introduced into the backbone network, the model can explicitly model the spatial position relationship of the feature map, enhance the spatial position perception capability of the personnel (especially small targets and blocked targets) in the construction scene under complex background interference, solve the problem that the traditional convolutional network is not sensitive to absolute position information, and improve the feature extraction accuracy in a complex background.

The present invention discloses a secure hierarchical federated learning method based on isolation forest anomaly detection, and relates to the technical field of electronic digital data processing. The method comprises the following steps: establishing a GRU self-attention model on the edge server side, extracting the temporal feature state of the global parameter through the GRU self-attention model, using the isolation forest algorithm to build a tree of the temporal feature state of the aggregated GRU self-attention model update parameter, comparing the anomaly detection strength evaluation coefficient with the anomaly detection strength threshold after the GRU self-attention model parameter is aggregated, and taking measures to adaptively adjust the anomaly detection strength or issue an alarm according to the comparison result. The present invention generates a personalized anomaly detection strength threshold decision for each server, fully considers the difference in the degree of attack suffered by different servers, and reasonably allocates the detection strength, thereby achieving the effect of improving the security of the server and solving the problem of insufficient server security in the prior art.

The invention relates to the technical field of remote sensing image processing and deep learning, and provides an improved Unet-based remote sensing image high-cleanliness water body extraction method, which comprises the following steps of: firstly, constructing a terrain embedding vector or a terrain guide map for terrain data (DEM and derived factors), and dynamically generating convolution kernel parameters through a multi-layer perceptron according to the terrain embedding vector or the terrain guide map; terrain adaptive feature extraction is realized in an encoder and a decoder through dynamic convolution; meanwhile, a terrain perception attention module is constructed, space attention and terrain guidance information are subjected to gating fusion to highlight a key area, and in the model training stage, joint loss including cross entropy, Dice and boundary weighting is adopted, and multi-scale deep supervision is introduced. Compared with a baseline network, the method provided by the invention has the advantage that the connectivity of F1, IoU and a narrow river channel is obviously improved on complex topographic data such as plateau and cross-cut mountain areas.

The application discloses a training method of a feature extraction model, an image processing method, equipment and a medium. The training method comprises the following steps: obtaining a sample image; performing feature extraction on the sample image by using a student feature extraction model and a teacher feature extraction model respectively, and obtaining multiple initial feature maps of different sizes, wherein the complexity of the student feature extraction model is less than that of the teacher feature extraction model; performing feature fusion on the multiple initial feature maps by using the student feature extraction model and the teacher feature extraction model respectively, and obtaining target feature maps corresponding to the initial feature maps, wherein the number of the target feature maps is the same as that of the initial feature maps; and adjusting parameters in the student feature extraction model according to the similarity loss between the multiple target feature maps obtained by using the student feature extraction model and the multiple target feature maps obtained by using the teacher feature extraction model. The above scheme can improve the accuracy of the student feature extraction model in the subsequent image processing process.

The invention relates to an insulation state evaluation method and system based on space charge characteristics, and the method comprises the steps: obtaining an original image of space charge density distribution, extracting a curve contour, and recognizing a curve type; based on the curve contour and the curve category, extracting a local average change rate and a peak value feature of the curve; and on the basis of the local average change rate and the peak value characteristics, in combination with the constructed sample database, performing insulation state evaluation in a weighted form to obtain an insulation state evaluation result. According to the invention, by introducing an image recognition technology, key characteristic quantities in a space charge distribution diagram can be extracted for insulating material samples in different aging states, and objective analysis and quantitative evaluation of space charge information are realized.

The present invention relates to the field of image super-resolution technology, and discloses a feature-enhanced cross-window lightweight super-resolution method, system and electronic device. The method comprises the following steps: constructing a self-attention mechanism based on Token dictionary intersection, and introducing a Token grouping mechanism to form a dictionary intersection self-attention module based on Token grouping, i.e., a TDAB module; constructing a feature-enhanced cross-window lightweight super-resolution network based on the TDAB module; and reconstructing a low-resolution image to be reconstructed using the feature-enhanced cross-window lightweight super-resolution network to obtain a reconstruction result. Based on the classical super-resolution method, the method introduces a self-attention mechanism based on Token dictionary intersection and a hybrid attention module based on Token grouping, which makes the model lighter while enhancing the model's ability to capture cross-window dependencies and the reconstruction effect of texture detail features.

The application relates to a power load prediction method based on a CRKformer model, electronic equipment and a storage medium, which comprises the following steps: acquiring time series data, constructing a data set and performing normalization preprocessing; converting the standardized data to a frequency domain, obtaining embedded features through linear dense operation, performing context shaping filter calculation, and obtaining a time domain optimization feature sequence through inverse Fourier transform; obtaining time series dependent features through the combination of self-attention mechanism and cross-attention mechanism of endogenous variable sequence embedding in the strong position coding weight; fitting a nonlinear relationship through a Kolmogorov-Arnold network, and enhancing the time series dependent features by using a learnable activation function; generating a prediction result by linearly projecting the endogenous output embedding, performing inverse normalization, and outputting a load prediction value. Compared with the prior art, the application has the advantages of inhibiting noise interference, accurately capturing time series dependence, and high prediction accuracy.

The present application relates to the technical field of medical image processing and artificial intelligence cross-application, in particular to a medical image detection method based on cross-channel and cell density adaptive mechanism. The method introduces a cell density adaptive attention module, a cross-channel feature enhancement module and a multi-magnification detection head on the basis of a YOLO main architecture, and constructs a multi-dimensional collaborative enhancement detection network model, so as to realize accurate detection and semantic feature strengthening representation of medical targets in different tissue cell density, different target scale and different image magnification scenarios. Through the core mechanisms of cell density adaptive modeling, cross-channel feature interaction and hierarchical feature fusion, the technical scheme effectively solves the technical problems of insufficient detection stability, weak model generalization ability and insufficient target semantic feature expression in the existing medical image detection technology, and finally realizes high-precision, low-delay and highly-scalable pathological image automatic detection.

The invention belongs to the technical field of elastic backing plate appearance defect detection, and particularly relates to a railway fastener elastic backing plate defect detection method based on improved YOLOv8. The method comprises the following steps: S1, data processing; s2, constructing a model; s3, model training; s4, defect detection; the model is improved by taking YOLOv8x as a basic algorithm, an improved YOLOv8 model is constructed, and the improvement comprises the step of replacing all C2f modules in the YOLOv8 with a GELAN module; in a top-down path of the Neck part, an improved SAFM module is embedded before up-sampling operation of each layer; and a normalized Wasserstein distance NWD loss function is adopted as a bounding box regression loss function. Compared with a conventional YOLO algorithm, the detection method provided by the invention has the advantages that the detection precision is remarkably improved, and intelligent and automatic online detection is realized.

The invention discloses a fabric density measurement system and method based on moire, and the system comprises a self-adaptive fabric detection module which preprocesses an original fabric image into a plurality of image blocks, constructs a demosaicking feature map according to the demosaicking feature indexes of the image blocks, generates a fabric region image, and transmits the fabric region image to a fabric density measurement module; positioning a fabric area containing moire patterns in the image; the moire feature extraction module is used for partitioning and enhancing the cloth area image to generate moire feature matrixes of warp dimension and weft dimension; and the density measurement module is used for verifying the effectiveness of the moire pattern characteristic matrix and realizing cloth density measurement. According to the method, the density of the cloth is measured by extracting the characteristics of the moire patterns generated by overlapping the shot cloth texture and the camera color filtering array, so that the problems of short deployment distance and small field of view of the existing fabric density measurement technology are solved.