5595results about How to "Improve features" patented technology

The invention provides an ultrahigh-resolution remote sensing image segmentation method based on frequency domain information fusion. The ultrahigh-resolution remote sensing image segmentation method based on frequency domain information fusion comprises the following steps: step 1, acquiring a Vaihingen and Potsdam remote sensing data set; step 2, building an ultrahigh resolution remote sensing image segmentation network fused with multivariate frequency domain information; step 3, inputting the Vaihinge training set into the ultra-high resolution remote sensing image segmentation network for training to obtain an optimal parameter model; 4, inputting the Vaihinge test set into the optimal parameter model, and outputting an accurate segmentation image of the remote sensing image; and 5, replacing the data set with a Potsdam training set, a Potsdam verification set and a Potsdam test set, and repeating the operations from the step 3 to the step 4. According to the model, the LKResNet-18 network is introduced, so that the receptive field and the feature representation capability of local branches are improved.

The invention relates to the technical field of sensor detection, and discloses a detection method and device based on a multi-mode gas sensor, and the method comprises the following steps: carrying out blending processing on a metal oxide semiconductor and a conductive material to obtain a bimodal nano composite material, and creating a multi-mode gas detection module, the preprocessed multi-mode signal data is acquired and obtained; acquiring and analyzing data of the temperature sensor and the humidity sensor to obtain environmental factor compensation parameters; performing adaptive calibration processing on the preprocessed multi-modal signal data according to the environmental factor compensation parameters to obtain calibrated multi-modal signal data; constructing a multi-input neural network model based on the calibrated multi-modal signal data, and training to obtain a gas recognition model; multi-mode signals collected and calibrated in real time are analyzed and judged based on the gas recognition model, gas type and concentration information is obtained, and high-precision and high-reliability gas detection is achieved.

The invention discloses a lightweight industrial product surface defect detection method, and belongs to the field of computer vision and artificial intelligence application, and the method comprises the steps: obtaining an industrial product surface defect image data set, and carrying out the preprocessing of the data; the method comprises the following steps: constructing a YOLO-IDL industrial product surface defect detection model, designing a lightweight hierarchical multi-scale feature extraction network as a backbone network, designing a scale rhinoceros neck as a neck network, and adopting a YOLOv8 head network as a head network; and training a YOLO-IDL industrial product surface defect detection model by using the preprocessed industrial product surface defect image data set to obtain a well-trained YOLO-IDL industrial product surface defect detection model, and deploying the well-trained YOLO-IDL industrial product surface defect detection model so as to detect industrial product surface defects and output defect types and positions.

The invention discloses a capture probe group for detecting pathogenic microorganisms and application of the capture probe group. The nucleic acid sequence of the capture probe group comprises a sequence as shown in SEQ ID NO. 1 to SEQ ID NO. 101. The capture probe group can cover and detect 119 types of bacteria, 32 types of fungi, 41 types of DNA viruses and 71 types of RNA viruses and covers common infectious disease pathogens, the kit containing the capture probe can cover and detect 88 types of drug-resistant and virulence genes, the detection performance of pathogenic microorganisms is greatly improved, and the sensitivity and specificity are improved.

The invention discloses an emotion-enhanced digital human driving and presenting system and method, and belongs to the technical field of computers.The system comprises an audio decoupling module, a potential diffusion model module, a human motion prior module, a training module and a rendered image generation module; predicting and reconstructing the facial expression and gesture action sequence to obtain a final action sequence; and generating a neural rendering image based on the final action sequence. According to the method, through voice input, emotion and style features in voice are captured, a cross attention feature fusion mechanism is introduced to enhance expression of emotion and personal style features, a voice-driven gesture generation task is expanded to a facial expression and gesture action synchronous generation task, the integrity of generated digital human actions is realized, and the user experience is improved. The emotional expressive force of facial expressions and gestures of the digital human is greatly improved.

The invention discloses an intelligent customer service method for adaptively adjusting an interaction strategy by using reinforcement learning, and the method comprises the following steps: S1, obtaining a natural language text inputted by a user, and generating a multi-dimensional state vector; s2, grouping the users by using a real-time dynamic clustering algorithm, and updating a clustering result; s3, generating an action space in a specific state, and enriching the action space in combination with a rule-based strategy and a learning-based strategy; s4, aggregating neighbor node features through a graph convolutional network, and extracting local features; s5, weighting adjacent node features by using a graph attention network, and dynamically allocating attention weights; s6, carrying out interaction strategy optimization by adopting a spiral ladder type deep exploration network, and optimizing action selection; and S7, combining online learning with an experience playback mechanism, continuously updating an interaction strategy, and making a quick response to dynamically changing user demands. According to the method, reinforcement learning, the graph neural network and the sentiment analysis technology are combined, and intelligent customer service interaction strategy self-adaptive adjustment and optimization are achieved.

The invention relates to the technical field of industrial internet, in particular to a smart energy management method and system based on the industrial internet, and the method comprises the following steps: S1, data collection: collecting energy consumption data and environment data in a production process in real time; s2, data fusion: carrying out fusion processing on multi-source data by adopting a data fusion technology; s3, distributed energy optimization: optimizing the energy consumption of each node on the distributed nodes in real time by using an edge computing technology; s4, intelligent load balance optimization: dynamically adjusting the operation load of each industrial production device; s5, multi-level fault prediction: performing fault prediction on the industrial production equipment; and S6, control strategy generation: generating a self-adaptive control strategy in combination with the fault prediction result of the industrial production equipment. According to the method, the lowest energy consumption of the optimization result is ensured, the load is uniformly distributed, local overload and energy waste are avoided, and the overall energy utilization efficiency is improved.

The invention provides a landslide displacement space-time prediction method, device and system and a storage medium, and relates to the technical field of engineering geology, and the landslide displacement space-time prediction method comprises the steps: obtaining the historical landslide multi-field time sequence monitoring data of each monitoring point of a to-be-measured region, constructing a monitoring point bitmap based on the historical space related data, constructing a time sequence characteristic matrix according to the historical displacement time sequence, the historical landslide external disturbance factor time sequence and the monitoring point bitmap; inputting the time sequence characteristic matrix and the monitoring point bitmap into a trained landslide displacement space-time prediction model to obtain a landslide displacement space-time prediction result; the landslide displacement space-time prediction model is constructed based on an improved GCN network and an improved BiLSTM network; according to the method, the space-time prediction precision of landslide displacement can be remarkably improved by combining the space correlation data and the external disturbance factors and utilizing the capability of the deep learning model.

The invention discloses a target detection method and device based on cross attention multi-scale fusion and a medium, and relates to the field of target detection, and the method comprises the steps: obtaining a visible light image and an infrared image of a to-be-detected target, and carrying out the target detection through employing a trained cross attention multi-scale fusion detection network; wherein the cross attention multi-scale fusion detection network comprises a backbone network, a background sensing module, a cross attention fusion module and a detection head; the backbone network is used for extracting features of different scales; the background sensing module is used for predicting illumination and contrast; the cross attention fusion module is used for performing cross fusion on the features extracted by the backbone network by adopting differential attention and homogeneous attention according to the illumination and contrast predicted by the background sensing module to obtain a multi-modal fusion feature map; and the detection head is used for determining the position and category of the to-be-detected target according to the multi-modal fusion feature map. According to the invention, the accuracy of target detection is improved.

The invention discloses a method for enhancing mRNA-LNP tissue targeting delivery. The invention provides a method for enhancing tissue targeting delivery of nucleic acid lipid nanoparticles, which comprises the following steps: modifying the surfaces of the nucleic acid lipid nanoparticles with tissue targeting with antibodies to obtain the nucleic acid lipid nanoparticles with modified antibodies and tissue targeting, the tissue targeting delivery property of the nucleic acid lipid nanoparticles is enhanced. On the basis of novel lung targeting LNP, an antibody modification strategy (double targeting) is combined, more efficient, specific and safer lung targeting mRNA delivery is successfully achieved, and the strategy is suitable for any type of targeting LNP (charge-driven DOTAP LNP and structure-driven STAR LNP in the project).

The invention relates to the technical field of intelligent temperature control, and discloses an intelligent temperature control lead-free tin bar production method and device.The method comprises the steps that data of a plurality of sensors of a production line are collected and preprocessed, and a time sequence data set is obtained; carrying out attention-guided position-sensitive multiple interpolation to obtain a target data set; constructing and training based on the target data set to obtain an anomaly recognition model; according to normal production data screened by the anomaly recognition model, analysis is carried out based on an implicit subspace recognition regression neural network of orthogonal basis decomposition and reconstruction, and a production process parameter model is obtained; based on the production process parameter model, a multi-task feature fusion framework is used for parameter optimization, target production process parameters are obtained, the problem of data missing in the production process can be effectively solved, abnormal conditions can be accurately recognized, self-adaptive process parameter optimization can be conducted according to the characteristics of different production lines or batches, and the production process parameter optimization efficiency is improved. Therefore, the production quality and efficiency of the lead-free tin bar are improved.

The invention belongs to the technical field of catalyst preparation and wastewater treatment, and particularly relates to a method for in-situ extraction of a high-activity monatomic catalyst from MOFs derived metal oxides and application. Preparing an N-coordinated metal organic framework from transition metal salt and an organic ligand through a precipitation method; then pyrolyzing at 300 to 800 DEG C to obtain a metal monatomic catalyst loaded on the metal oxide and the carbon-nitrogen composite carrier; the organic ligand comprises trimesic acid and a nitrogen-containing ligand. According to the catalyst, the carrier effect of metal oxide and carbon is coupled, in-situ regulation of a monatomic metal site coordination environment is realized by introducing a nitrogen-containing ligand, the efficiency of catalyzing and activating persulfate to degrade phenol is effectively improved, and meanwhile, good anti-interference and cyclic utilization characteristics are shown. The catalyst disclosed by the invention is simple in preparation process, excellent in catalytic effect, high in stability, low in requirements on a treatment process and treatment equipment and free of secondary pollution, and has important significance in the field of wastewater treatment.

The invention discloses a metal surface defect detection method. The method comprises the following steps: establishing a metal surface defect image data set; the metal surface defect image is preprocessed, and it is guaranteed that multiple defect types exist in the same image; a YOLOv7 network model is improved; carrying out training evaluation on the metal surface defect data set by using the improved YOLOv7 model, and selecting proper hyper-parameters in the training process; data obtained after model training are analyzed and judged, and the model is adjusted and optimized; and keeping the optimal model weight, and carrying out experimental testing on the trained model. The method comprises the following steps: establishing a multi-type metal surface defect data set; a polarization self-attention mechanism and an SPDConv module are introduced into the neck network, and important information is captured in a self-adaptive mode while image information is reserved; and an NWD loss function is introduced, and the positioning capability of the network on small-size defects is improved. The method can meet the requirements of metal surface defect detection under different working conditions, and has good robustness.

The invention discloses a symptom management and prognosis evaluation method and system for a tumor radiotherapy patient, and relates to the technical field of tumor radiotherapy patient management.The method comprises the steps that basic information of the tumor radiotherapy patient is obtained, prognosis evaluation is conducted according to the basic information, a first result is generated, and the risk category is judged according to the first result; obtaining symptom information of a tumor radiotherapy patient, evaluating the symptom severity according to the symptom information, and generating a second result; and performing symptom management according to the second result and the risk category, generating prognosis data, and updating the first result through the prognosis data. According to the invention, by fusing the dual contrast learning network and the dual-path decision tree integrated model, the dynamic association analysis and adaptive feature enhancement of the multi-dimensional medical data of the tumor radiotherapy patient are realized, so that a precise prognosis evaluation and a personalized symptom management scheme are provided for clinical practice.

The invention relates to the technical field of data analysis, and discloses a data analysis method based on a chain operation benefit supervision system, and the method comprises the steps: carrying out the collection and preprocessing of the multi-dimensional data of each store of a chain operation enterprise, and obtaining the multi-dimensional time series data; performing multi-dimensional similarity calculation and fuzzy clustering to obtain a store clustering result and a fuzzy membership degree; constructing a decision tree model and performing store operation state evaluation to obtain a store operation state evaluation result; time sequence feature extraction is carried out, a semantic segmentation network is constructed, and a store business prediction model is obtained; performing deep prediction on the multi-store business data to obtain a store business prediction result; and based on the store operation state evaluation result, the store business prediction result and the fuzzy membership degree, comprehensive anomaly prediction and resource optimization configuration are carried out to obtain an anomaly prediction report and an optimization scheme, so that the accuracy of operation state evaluation is improved.

The invention relates to the technical field of molecular markers, and discloses an SNP (Single Nucleotide Polymorphism) molecular marker related to the litter size character of pigs and application of the SNP molecular marker. The SNP molecular marker is located at a 34977540bp site of a chromosome 2 of an international pig reference genome Sscrofa11.1 version, and the nucleotide sequence of the site is G or A and is located in an exon region of an LUZP2 gene. The invention also discloses an application of the SNP molecular marker in breeding of the litter size character of the boar, and provides a method for breeding or auxiliary breeding of the boar with high / low litter size based on the marker. The invention also provides a specific primer for identifying the SNP site. By applying the technologies, accurate prediction and efficient screening of the litter size of the large white pigs can be achieved, the breeding efficiency of excellent breeding pigs is remarkably improved, and the population improvement process is accelerated. The invention also provides important technical support for sustainable development of pig raising industry, also provides theoretical basis for research in related fields, and is expected to promote further development of pig genetic improvement work.

The invention relates to the technical field of behavior recognition, in particular to a machine learning-based surgical personnel video behavior recognition method and system. The method comprises the following steps: acquiring a video frame sequence and a point cloud frame sequence from a hand washing area of an operator through an RGB camera and laser ranging equipment to obtain the video frame sequence and the point cloud frame sequence; performing hand point cloud segmentation on the point cloud frame sequence to obtain hand point cloud data; performing action original data generation according to the video frame sequence and the hand point cloud data to obtain action original data; and extracting position and attitude information of hand key points from the original motion data to obtain video key point feature data. Visual information is captured through the RGB camera, three-dimensional depth information is obtained through the laser ranging device, fine analysis and evaluation are carried out on actions in combination with the deep learning network, and finally higher recognition precision, robustness and automation level are achieved.

The invention relates to the technical field of genomics, in particular to a gene detection data cleaning method and system based on artificial intelligence. The method comprises the following steps: acquiring gene detection data, and performing preliminary filtering on the gene detection data based on removal of low-quality reads to obtain preliminary filtering data; performing linker sequence identification and removal on the preliminarily filtered data, and performing sequencing error correction to generate gene sequence data; performing GC content analysis and repetitive sequence identification on the gene sequence data to obtain sequence feature data; performing dynamic quality threshold adjustment according to the sequence feature data, and performing secondary filtering on the gene sequence data to generate high-quality sequence data; and performing sequence pattern recognition on the high-quality sequence data to obtain sequence pattern feature data. According to the method, deep analysis is carried out by utilizing sequence characteristics such as GC content and nucleotide distribution, and the recognition capability on sequence complexity and variation modes is improved.

The invention relates to the technical field of medical diagnosis systems, in particular to an early gonitis diagnosis system and method based on multi-modal deep learning, a data collection module is used for collecting and preprocessing multi-modal data including knee joint X-ray image data, MRI (Magnetic Resonance Imaging) image data and clinical evaluation data; the model construction module constructs a deep learning model by combining a convolutional neural network and a Transform architecture, realizes mapping from multi-modal data to a clinical assessment phenotype label by adopting a feature automatic learning module and a potential relationship mining module, and integrates a Grad-CAM interpretability method for understanding a model prediction basis; diagnostic data are continuously collected to be fed back to the data collection module for model optimization, the convolutional network is good at capturing local image features, the Transform architecture is good at modeling a long-range dependency relationship, and the feature extraction and relationship modeling capabilities of the model are greatly improved through combination of the convolutional network and the Transform architecture.

The invention discloses a copper alloy pipe and a preparation method and application thereof, and the copper alloy pipe comprises the following components in percentage by mass: 0.01%-0.80% of Sn, 0.005%-0.05% of P, 0.01%-0.10% of Zr, 0.001%-0.01% of B, 0.001%-0.01% of Mn and the balance of Cu and inevitable impurities. Wherein the copper alloy pipe contains MnZr and BMn compound particles, the number of the MnZr compound particles on the cross section of the copper alloy pipe is 0.005-0.15 / mu m < 2 >, and the number of the BMn compound particles on the cross section of the copper alloy pipe is 0.005-0.2 / mu m < 2 >. By adding Zr, B and Mn and controlling the content of Zr, B and Mn, the preparation process and the like, the number and uniformity of MnZr and MnB compounds are regulated and controlled, and the tensile strength, ductility, corrosion resistance, bending, flaring, welding, pressure resistance and other machining performance of the copper alloy pipe are improved; and the prepared copper alloy pipe fully meets the application requirements of heat conduction pipes of air conditioner heat exchangers, air conditioner indoor and outdoor unit connecting piping and the like.

The invention discloses a preparation method of PEDOT: PSS (poly (3, 4-ethylenedioxythiophene): poly (styrenesulfonate))-based composite fiber with silver nanoparticles loaded on the surface. The method comprises the following steps: uniformly mixing a silver nanowire solution with a PEDOT: PSS solution, or uniformly mixing the silver nanowire solution, the PEDOT: PSS solution and a polymer solution to obtain a spinning solution; extruding the spinning solution into a coagulating bath through a wet spinning process and soaking to obtain nascent composite fibers; and drying the nascent composite fiber to obtain the PEDOT: PSS-based composite fiber of which the surface is loaded with silver nanoparticles and the interior contains silver nanowires. The interior and surface structures of the PEDOT: PSS fiber are simultaneously modified, the interior of the fiber contains a silver nanowire structure, and the surface of the fiber has a silver nanoparticle structure. The composite fiber has the characteristics of ultra-sensitive SERS sensing capability, high conductivity, electric heating, molten drop resistance and antibacterial property.

The invention relates to integrated airborne communication antenna equipment and a communication method. The method comprises the following steps: carrying out CSI data acquisition and space-time focus modulation on a communication channel to obtain CSI space-time characteristic data; performing residual vector quantization operation to generate a quantized data stream, and performing recursive coding processing to obtain a target data stream; mapping the target data stream to a three-dimensional tensor space, constructing a real value tensor model, and performing iterative operation to obtain a CSI feature vector; performing fractional order wavelet transform decomposition on the CSI feature vector to obtain a radio frequency feature set, and inputting the radio frequency feature set into a federal fractional scattering network for feature processing to obtain an identity authentication result; and calculating an amplitude-phase weight coefficient of the antenna array, generating a beam parameter, and performing compensation matrix operation on the beam parameter according to the CSI feature vector to obtain a pre-coded signal. According to the invention, the dynamic adjustment of the antenna beam parameters is realized, and the accuracy and anti-interference capability of beam forming are improved.

The invention belongs to the technical field of intelligent transportation, and discloses a driving state recognition method and system based on multi-modal fusion. According to the method, the wrist movement information of the driver is collected through the non-invasive physiological bracelet, the wrist movement information is combined with traditional physiological signals, driving performance data and the like, and the accuracy and robustness of driving state recognition are improved. The invention provides a feature dynamic weighting model, the model captures a long-term dependency relationship in a time sequence by using an LSTM module, and weights the features among different modes by using a cross attention mechanism module, so that the dynamic importance of each feature can be adjusted according to different driving states, and the accuracy of driving is improved. And deep interaction and feature enhancement of different modal data are realized by using a Transform module. According to the method, the XGBoost model is utilized, driving state classification is carried out based on the weighted multi-modal features, and different driving states can be accurately distinguished.

The invention belongs to the technical field of image processing, and discloses a transformer fault detection method based on an improved YOLOv8 model. Aiming at the problems of inaccuracy in feature extraction, low precision, poor real-time detection capability and the like existing in fault detection of a traditional transformer fault detection method, the invention establishes a transformer fault detection model based on an improved YOLOV8 model architecture, a self-calibration convolution defogging module is introduced, a generative adversarial network (GAN) and a multi-source data fusion network structure are added, a BiFPN structure is used, an ASPP module is replaced by an ASPFCSPC module, and a YOLOV8 model architecture is improved, so that the feature extraction capability of a transformer fault detection model is enhanced, the calculation complexity of the model is reduced, and the fault detection efficiency of the transformer is improved. The transformer fault detection precision, accuracy and real-time detection capability are improved, and the stability and adaptability of the model are enhanced.

The invention relates to the technical field of side slope monitoring and early warning, in particular to a mine site side slope remote online real-time monitoring and early warning system which comprises a multi-source data monitoring unit, a data processing and analyzing unit, a side slope instability judging unit, an instability grading early warning unit and an early warning verification pushing unit. The method comprises the following steps: collecting monitoring data of a plurality of data sources to obtain time sequence monitoring data of each data source, carrying out noise reduction on the time sequence monitoring data, then decomposing the time sequence monitoring data, calculating a time-frequency feature vector through an energy density function, and optimizing the time-frequency feature vector through feature screening; calculating a slope stability evaluation index by optimizing the time-frequency feature vector, constructing an instability risk index to judge a slope instability risk level, constructing an early warning index, and combining the early warning index with the slope instability risk level to establish a graded early warning mechanism to obtain a slope instability early warning level; and verifying the slope instability early warning grade, and performing early warning according to the slope instability early warning grade passing the verification. According to the invention, the false alarm rate of slope monitoring and early warning can be reduced.

The invention discloses a rainy day target detection method based on paired image data generation and knowledge distillation, and belongs to the technical field of automatic driving target detection. The method comprises the following steps: acquiring a real image of a road scene and a semantic segmentation mask; inputting the semantic segmentation mask and the weather condition text information into a trained improved diffusion model to obtain a road scene generation image under a rainy day condition; introducing a CoT module and an ODConv module into the target detection model for optimization, and then improving the target detection model by adopting a dynamic weighted aggregation loss function to obtain an optimized target detection model; performing knowledge distillation to obtain a rainy day target detection enhancement model; and performing target detection by using the rainy day target detection enhancement model. The method provided by the invention is excellent in performance in a rainy day scene, has good generalization ability, can provide higher-precision target detection in various complex scenes, and effectively improves the detection accuracy and stability.

The invention discloses a ginsenoside Rg3 nano-liposome easier for transdermal absorption as well as a preparation method and application thereof, and belongs to the technical field of liposome preparation. The ginsenoside Rg3 nano-liposome easier to transdermally absorb, provided by the invention, is prepared from the following raw materials in parts by mass: 5 to 20 parts of polyhydric alcohol, 2 to 10 parts of ginsenoside Rg3, 1 to 8 parts of phospholipid, 1 to 5 parts of cholesterol, 1 to 2 parts of resveratrol, 0 to 2 parts of a compound functional substance and 1 to 5 parts of water. The ginsenoside Rg3 nano-liposome easier to transdermally absorb, provided by the invention, has excellent anti-oxidation and anti-inflammatory characteristics through a compound synergistic effect of all the components, can regulate the skin state, comprehensively protect the skin from being damaged by free radicals and inflammation, solves the skin problem caused by external stimulation, and has a good application prospect. The skin care composition has the effects of protecting skin, maintaining skin barrier function, improving skin redness and improving skin stability, can be compounded with any skin care preparation for use, and has wide market application prospect.

The invention relates to an HPV16 type E6 oncoprotein monoclonal antibody as well as a preparation method and application thereof. The monoclonal antibody specifically bound with the HPV16E6 recombinant antigen is obtained through screening, the monoclonal antibody is applied to a biotin-avidin amplification ELISA system, the E6 oncoprotein of HPV16 can be specifically detected, the sensitivity is 100 pg / mL, the monoclonal antibody does not intersect with oncoprotein of other high-risk HPV, and the E6 oncoprotein of HPV16 can be specifically detected. High specificity and sensitivity are shown for cervical exfoliated cell sample detection of normal people, CIN3 people and cervical cancer people, and the method can be applied to early screening of cervical cancer in the future; compared with traditional exfoliative cytology examination, the method is faster and more accurate, meanwhile, the risk of false positive and misdiagnosis caused by too high sensitivity of nucleic acid detection can be avoided, and the method has very high application value.

The invention discloses an abnormal behavior detection method and device based on improved YOLOv8 and a storage medium. According to the method, a new FDPN structure is provided to replace a neck network architecture of an original YOLOv8 model; the FDPN has the advantages of a feature focusing module and a feature diffusion mechanism, the feature focusing module effectively captures and integrates cross-scale feature information through parallel deep convolution operation, and the diffusion mechanism promotes contextual information to flow among scales, so that each feature map contains detailed semantic information, the small target detection precision is improved, and the detection efficiency is improved. And the adaptability of the network to shielding and complex backgrounds is enhanced. In addition, an RCSOSA module and a DyHead dynamic prediction head are also introduced; the RCSOSA module extracts and enhances diversified features through a RepVGG module, optimizes feature selection and channel shuffling through a plurality of RCS modules, and enhances feature fusion between channels; and by using the DyHead dynamic prediction head, the recognition capability of the shielded crowd can be improved. According to the method, the robustness and the precision of abnormal behavior detection in a complex scene are remarkably improved.

The invention belongs to the technical field of computer vision and pattern recognition, and particularly relates to a clothes changing pedestrian re-recognition method based on clothes perception mixing and foreground enhancement. The method comprises the following steps: acquiring image data to construct a data set, performing text description on acquired images, preprocessing the image data in the data set, and dividing the image data into a training set and a test set; a clothes changing pedestrian re-recognition model is constructed, the model comprises a text feature extraction unit, an image feature extraction unit and a feature fusion unit, the pedestrian images in the training set are input into the clothes changing pedestrian re-recognition model for training, and a loss function of each unit is obtained; summarizing the loss to obtain total loss, and training and optimizing the clothes changing pedestrian re-identification model through the total loss; inputting data in the test set into the optimized and trained clothes changing pedestrian re-identification model for similarity comparison, and selecting an image with the highest similarity as a final retrieval result; according to the invention, the accuracy of re-identification of clothes changing pedestrians can be improved.