134 results about "Linear classifier" patented technology

The invention provides a multi-modal speech emotion recognition method based on a pre-training model, and the method comprises the following steps: 1, obtaining a data set which comprises a plurality of data pairs, each data pair comprises corresponding text data and audio data, and each data pair is provided with an emotion label; 2, acquiring text semantic features from the text data, and acquiring audio acoustic features and voice emotion features from the audio data in the same data pair; 3, performing feature fusion on the text semantic features, the audio acoustic features and the voice emotion features to obtain fusion features; and step 4, inputting the fusion features into a linear classifier, outputting the fusion features as scores of each emotion after the fusion features pass through a softmax layer, and taking the emotion corresponding to the emotion label corresponding to the maximum score as a recognition result. According to the method, a voice emotion-oriented fusion strategy is provided based on an attention mechanism, and the prediction accuracy of the voice emotion is remarkably improved.

The invention provides a landslide classification method and system based on a visual language model and a cross attention mechanism. The method and the system specifically comprise the following steps: data preprocessing: carrying out Canny edge detection on an RGB image, and calculating terrain attributes such as a gradient and a slope direction for a DEM (Digital Elevation Model); feature extraction: capturing local features by adopting a reflection filling convolution layer and multi-scale residual connection; a visual language model is introduced, wherein semantic enhancement features are extracted through image-text alignment by means of the visual language model; cross self-attention fusion: capturing a global context through self-attention, and focusing heterogenous data complementary information by cross attention; and classifying and outputting: outputting a result by using global average pooling and a linear classifier. Through the visual language model and the cross self-attention mechanism, the landslide recognition capability under the complex terrain is effectively improved, an efficient and reliable technical means is provided for geological disaster monitoring, and the method can be widely applied to the fields of landslide recognition, risk assessment and the like.

The invention belongs to the technical field of electric digital data processing, and more specifically relates to an agricultural machinery behavior identification method fusing interpolation enhancement and depth spatial-temporal feature modeling. The method comprises the following steps: acquiring GNSS trajectory data of an agricultural machine and preprocessing the trajectory data; performing data enhancement on the road points in the data set by adopting a local interpolation method based on DBSCAN (Density Based Spatial Clustering of Applications with Noise) clustering; designing a set of spatial distribution feature extraction method; performing feature extraction on the trajectory data after data enhancement by combining motion feature extraction and spatial distribution feature extraction; introducing a variational auto-encoder VAE to carry out potential feature extraction on the initial features; then, a ResBiLSTM network is adopted to carry out space-time modeling; and finally, an end-to-end classification decision is realized through a linear classifier, so that intelligent identification of the operation behavior of the agricultural machine is realized. According to the method, two key problems in the existing track identification technology are solved, namely, data distribution is seriously unbalanced, and the spatial distribution characteristics of track points are often neglected.

The invention provides a multi-modal large model mental ability improving method and device based on attention intervention, and relates to the technical field of artificial intelligence. The method comprises the following steps: generating a multi-modal psychological theory data set; inputting the data to a multi-modal large language model to be evaluated, and constructing an input sequence of a Transform network; extracting an activation value of each attention head in each layer of attention network for each input sequence, and defining a belief label of the activation value from the perspective of a leading angle; taking the trained linear classifier as a probe model to obtain a predicted belief label, and constructing a sensitivity score of each attention head; and extracting an intervention target direction vector of each high-sensitivity attention head, and executing directional activation intervention on each high-sensitivity attention head to obtain a multi-modal large language model after the psychological theoretical ability is improved. According to the method, innovation is systematically carried out from three dimensions of data construction, mechanism detection in the model and reasoning intervention enhancement, and the key blank in the prior art is filled.

The invention discloses an image description generation system, a training method, a generation method and electronic equipment, and belongs to the technical field of image description. According to the method, visual features of an image are mapped to a visual and language comparable space, after a semantic information sequence is obtained, cross-modal semantic calculation of the semantic information sequence and the visual feature sequence is achieved through a Transform decoder, the middle hidden state of each candidate vocabulary is obtained, and then a corresponding directed acyclic graph is constructed; and after an optimal path is selected from the directed acyclic graph, the optimal path is directly mapped into image text description by a linear classifier. According to the method, visual information of the image and semantic information contained in the image are fully utilized, the sequential relation between words is learned by introducing the directed acyclic graph, the smoothness of description generation is improved, the method has non-autoregressive decoding attributes, and high-quality image text description can be generated at a high speed.

The invention relates to a medical image processing technology, and discloses a pan cancer species histopathologic full-slice classification method and device, which is a trained multi-modal pan cancer species classification model based on histopathologic full-slice image data and full-slice diagnosis report pairing data of various tissue parts and various cancer species, and has the advantages of wide application range and high efficiency. And full slices of various pathological tissues can be classified. For classification tasks of different tissue parts, only text prompts (prompts) are used, i.e., a candidate cancer species list is set, and zero-sample full-slice classification can be realized without training. And for rare cancer species, full-slice classification can be carried out. Besides, the model has high expansibility, in order to better adapt to downstream tasks, better classification performance can be achieved by combining slice embedding with a linear classifier for use, freezing a slice embedding part and only finely adjusting the linear classifier, and the requirement for the training data size is low.

A small sample modulation identification method based on combination of multi-domain contrast learning and reinforcement learning comprises the following steps: 1) obtaining a data set, dividing the data set into a base set, a support set and a verification set, converting the data set into a time domain, a frequency domain and a constellation diagram form, and obtaining a corresponding encoder and a common projection head; 2) inputting the base set data into an encoder for feature extraction, inputting the obtained features and accuracy as a state and an award into a decision body for reinforcement learning training to obtain a data enhancement action, inputting the original data and the enhancement data into the encoder and a projection head, and performing comparative learning training in a high-dimensional space to obtain a data enhancement action; the reinforcement learning training is repeated until the rewards are not increased any more; 3) loading and freezing encoder parameters, and training an attention module and a linear classifier by using a support set; and 4) loading and freezing all module parameters, and verifying the performance by using the verification set. According to the method, the signal model can have good performance in a small sample scene.

The invention relates to the field of medical auxiliary diagnosis, in particular to a state space duality multi-instance pathological image classification method and system, and the method comprises the following steps: obtaining a lung cancer pathological full-slice image to be detected based on a pathological database; obtaining a tissue area of the lung cancer pathological full-slice image, and cutting the tissue area to obtain an image block set; obtaining a feature vector set of the image block set by using a feature extractor, constructing a fusion feature set based on the feature vector set, and obtaining a weighted feature set according to the fusion feature set; performing dichotomy mapping on the weighted feature set according to a linear classifier to obtain a dichotomy prediction probability vector; and obtaining a cancer prediction category of the lung cancer pathological full-slice image according to the dichotomy prediction probability vector so as to classify the cancer type of the lung cancer pathological full-slice image. The method is used for automatic diagnosis of cancer pathological images and provides efficient and reliable technical support for medical auxiliary diagnosis.

The invention relates to the technical field of retina image quality evaluation, and particularly discloses a retina image quality evaluation method based on double-path frequency domain cross attention fusion, comprising the following steps: step S1, acquiring an RGB image; s2, preprocessing the quality of the BGR image through a contrast limit adaptive histogram equalization algorithm to obtain a CLAHE image; s3, inputting the RGB image and the CLAHE image into a dual-path frequency domain DenseNet121 network, and extracting features of the RGB image and the CLAHE image through a frequency domain attention mechanism algorithm; s4, refining the extracted features through a double cross attention algorithm, and obtaining integration of the BGR image and the CLAHE image; and S5, obtaining a quality evaluation result through a linear classifier. According to the method, important retina image structure details are reserved, and the adaptability of the model to different illumination conditions is also improved.

The embodiment of the invention provides a training method for a long-tail image recognition model, and belongs to the technical field of data processing, and the method specifically comprises the steps: sampling a picture through employing a balance sampling strategy and a random sampling strategy, and carrying out the data enhancement of a random sampling picture, and obtaining a first sample; executing the CamMix operation to obtain a second sample; inputting the first sample and the second sample into a backbone network to obtain sample features; obtaining cross entropy loss, and taking parameters in the linear classifier as prototype features which are not subjected to L2 normalization; enabling the prototype features and the sample features to pass through a multi-layer perceptron network to obtain L2 normalized features, and inputting the L2 normalized features into a prototype decoupling contrast learning loss function to obtain contrast loss; and obtaining a total loss function by combining the cross entropy loss and the comparison loss, and guiding backbone network training according to the total loss function to obtain a long-tail image recognition model. Through the scheme of the invention, the model identification accuracy and robustness are improved.

The invention discloses a text detection method and device, and relates to the technical field of data processing, and the method comprises the steps: carrying out the corner detection of text data of an original image based on a Harris corner detection algorithm, and obtaining a corner detection image; performing weighted fusion on the corner detection image and the original image based on a preset transparent parameter to obtain a fused image; performing feature extraction on the fused image based on a preset convolutional neural network model and a Vision Transform network model, and mapping the extracted features to a character probability based on a linear classifier and a softmax function; and generating extracted text data based on the character probability of each extracted feature. Compared with the prior art, corner detection is carried out on the text data of the original image based on the Harris corner detection algorithm to obtain the corner detection graph, even if the text data is rich and diversified in shape, the Harris corner detection algorithm can still carry out accurate positioning and extraction on key corners in the original image, and the accuracy of the corner detection graph is improved. Therefore, the method has better adaptive capacity for detecting the text data with abundant and diversified shapes.

The invention discloses an unmanned aerial vehicle detection method based on self-supervised learning. The method comprises the following steps: collecting CSI data in multiple scenes for a pre-training task of a Transform model; after knowledge in the pre-trained Transform model is migrated to a lightweight CNN model by using a knowledge distillation method, a linear classifier is added to the CNN model, and the CNN model is optimized by using a small amount of label data through quantitative perception fine tuning; and performing real-time unmanned aerial vehicle detection on the test data through the optimized CNN model. According to the method, channel state information in a scene where an unmanned aerial vehicle possibly exists is utilized, a Transform model is pre-trained through a two-dimensional reconstruction task, knowledge is migrated to a lightweight convolutional neural network through knowledge distillation, and then a quantitative perception fine tuning technology is used for training, so that efficient classification detection in multiple scenes is realized. The method has the advantages of high detection precision, low label dependence, light model weight, high cross-scene generalization capability and the like.

The invention discloses an unbalanced node classification method based on graph contrast learning, and belongs to the technical field of artificial intelligence. According to the method, a graph comparison learning framework of adaptive balance data is provided, minority classes can be automatically identified, the minority class performance is improved, and then the overall performance of the model is improved. Firstly, an Encoder-Decoder architecture is used for pre-training, and compared with a traditional pseudo tag generation method, an unbalance rate self-adaptive sampling strategy is designed, the unbalance rate of data is calculated according to pseudo tags, and the sampling strategy is selected in a self-adaptive mode. For a data set with a low unbalance rate, a simple downsampling method is adopted, and the proportion of minority class information is increased; for a data set with a relatively high unbalance rate, a mixed sampling strategy is adopted, and over-sampling and down-sampling are combined, so that the information loss of majority of nodes is reduced while the information proportion of minority of nodes is increased. In addition, the pre-training model used in the invention can provide more accurate label information, thereby improving the distinguishing ability of the model in subsequent GCL training. Then, a new data augmentation technology is designed, in the node masking process, pseudo label information is utilized, information of minority class nodes is reserved preferentially, and meanwhile majority class nodes are masked; the method is helpful for the model to better capture minority class features in an unbalanced data set. And finally, a linear classifier is used for classification. According to the method, the unbalanced node classification performance under the self-supervision condition can be effectively improved.

The invention discloses a handwritten text recognition method based on multi-stage enhancement. The handwritten text recognition method comprises the following steps: acquiring a handwritten text image; constructing a hierarchical dynamic multi-scale CNN backbone network to obtain a visual feature sequence; inputting the visual feature sequence into a time sequence multi-scale module to obtain a local enhanced feature sequence; performing global modeling on the local enhanced feature sequence by using a Transform encoder to obtain a global visual feature sequence; enhancing the global visual features to obtain a time sequence context feature sequence; dynamic weighted fusion is carried out on the global visual features and the time sequence context features through a gating fusion module; and sending the fused features into a linear classifier and a CTC decoder to obtain an identification result. According to the method, intelligent arbitration of global and time sequence features is realized through new technology application of a hierarchical multi-scale CNN trunk, time sequence context enhancement and a gating fusion mechanism, and the recognition accuracy and robustness are remarkably improved.

The invention discloses a tinnitus grading method and system based on EEG electroencephalogram analysis and deep learning, and relates to the technical field of tinnitus grading, and the method comprises the steps: in a pre-training stage, converting EEG signals of tinnitus patients with different tinnitus severity degrees into EEG images, inputting the EEG images into an EEG self-encoder, extracting potential features in the EEG signals, and carrying out the recognition of the potential features in the EEG signals; and inputting into a five-classification linear classifier in a downstream fine tuning stage to obtain the probability of each tinnitus severity level, and predicting the tinnitus severity according to the probability. Through training, a deep neural network model containing an EEG auto-encoder used for EEG feature extraction and a linear classifier used for generating the tinnitus severity of a patient is obtained. The probability that the tinnitus severity of the to-be-graded tinnitus patient belongs to each severity level in the five levels is obtained by using the deep neural network model, so that the tinnitus severity is accurately predicted, and medical personnel are helped to formulate a personalized rehabilitation scheme for the patient.

The invention discloses a multi-modal disease data processing system based on an interpretable model, and relates to the technical field of electrical digital data processing, and the system comprises a report extraction concept library which is formed by embedding a plurality of concepts; the pre-training image encoder is used for extracting feature vectors of the multi-modal radiology images to obtain image embedding; the multi-modal concept score calculation module is used for calculating according to the concept embedding in the image embedding and report extraction concept library to obtain a multi-modal concept score; the multi-modal concept score correction module is used for providing a window for modifying a multi-modal concept score; the linear classifier is used for calculating classification result data based on the multi-modal concept score; and the report generation module is used for generating a data processing classification report according to the classification result data and concept embedding of the report extraction concept library. According to the method, the practicability, reliability and efficiency of the deep learning model in medical image judgment are remarkably improved, and the trust and satisfaction of doctors and patients to the AI medical auxiliary system are enhanced.

The invention discloses a human body activity identification method and system capable of effectively capturing variable time-space relations in sensors and between the sensors. The method comprises the following steps: 1) carrying out cross-variable fusion on local cross-channel fusion features; 2) performing global time aggregation on the local cross-variable fusion features to obtain sensor feature data; 3) integrating different sensor feature data by using a self-attention mechanism to generate human body activity features; and 4) inputting the human body activity features into a full-connection linear classifier to obtain human body activity categories. The system comprises N wearable motion sensors, a data conversion module, a local time feature extraction module, a cross-channel fusion module, a cross-variable fusion module, a global time aggregation module, a human body activity feature extraction module and a human body activity classification module. By capturing the space-time relationship between the interior of the sensor and the sensors, the frame can provide more systematic information, so that the model is more accurate during analysis and prediction.

The application discloses a kind of light-weight human key point detection method and device based on model pruning, comprising the following steps: (1) the original image is convolved and continuously down-sampled, and different scale low-resolution images are obtained;(2) a multi-branch convolutional neural network is constructed, different resolution images are input into branch network respectively, and information fusion layer is set between different branches;(3) a linear classifier is added to each convolution block in the network, and the contribution of each convolution block is calculated using the accuracy of the linear classifier;(4) the contribution of the convolution block is sorted, and the convolution block with low contribution is deleted according to the pruning ratio to obtain a pruned network;(5) the pruned network and the original convolutional neural network are trained together using the knowledge distillation method;(6) the image to be detected is input into the trained pruned network, and the detection result of human key points is obtained.The application can effectively improve the accuracy and reduce the model parameters.

The invention discloses a network flow classification method and device based on multi-modal feature fusion, and the method comprises the steps: obtaining network flow data, extracting an IP address as a node, and constructing a communication graph; time modal features and event modal features are extracted for each node, and standardization processing is carried out on the time modal features and the event modal features; the time modal features and the event modal features are aligned; fusing the two types of aligned features; performing multi-layer neighbor sampling and hierarchical aggregation on the communication graph by using a graph neural network model to obtain structure-enhanced node features containing multi-hop neighbor information, inputting the node features into a linear classifier, outputting a category probability, selecting a category with the maximum probability as a prediction result, and performing iterative training to obtain a classification model; and mapping the classification model to a data plane of the programmable switch to realize online reasoning of the data packet. According to the method, high-precision identification of various network traffic types is realized by fusing the spatial modal, time and event modal characteristics.

The invention discloses a alumen ustum image clear area screening method based on superpixel segmentation and feature classification. The method comprises the following steps: generating a definition image of an original alumen ustum image by using a Laplace operator; enhancing the color contrast of the original alumen ustum image by using the definition image; segmenting the enhanced image into a plurality of small regions by using an SLIC superpixel segmentation method, and storing segmentation boundaries of all the small regions; applying the segmentation boundary to the original alumen ustum image, filling all small areas into a minimum enclosing rectangle, and storing the position and structure information of the minimum enclosing rectangle; using a ResNet feature extraction network to perform feature extraction on the small regions obtained by segmentation; and performing definition classification on all the small areas by using a linear classifier, and returning a classification result to an original image to obtain a clear alumen ustum image. According to the method, automatic identification and screening of clear alumen ustum structures in the image are realized through image small region division and region feature discrimination.

The present specification relates to the technical field of natural language processing and artificial intelligence, and specifically discloses a text multi-label classification method and device, wherein the method comprises: receiving a text multi-classification request; the target text data is carried in the multi-classification request; inputting the target text data into a pre-training model to obtain a target pre-training word vector matrix corresponding to the target text data; generating a prompt template matrix based on a pre-constructed adaptive prompt template; the adaptive prompt template is constructed through iterative training based on contrastive learning; splicing the prompt template matrix and the target pre-training word vector matrix to obtain a spliced target sentence vector representation; inputting the target sentence vector representation into an encoding model to obtain a target sentence representation corresponding to the target text data; and mapping and classifying the target sentence representation using a linear classifier to obtain a label set corresponding to the target text data. The above method can improve the accuracy and efficiency of multi-label classification.

Systems and methods for property guided molecular optimization using artificial intelligence diffusion models. An equivariant continuous denoising diffusion implicit model autoencoder framework (DDIM-AE) can be trained (510) on a conformational dataset to predict raw data from data corrupted by a time-dependent noise to obtain a trained DDIM-AE that ensures controlled generation of threedimensional (3D) molecules. Linear optimization of semantic embeddings of 3D molecules can be performed (520) with a linear classifier to achieve a target property value from desired properties and obtain an optimized embedding. An optimized 3D molecule that includes molecular conformation with the desired properties while preserving interactions with biochemical molecules can be generated (530) from the optimized embedding with the trained DDIM-AE.

The invention provides a cholecystectomy CVS evaluation system, method and device based on SMIL, and relates to the technical field of video processing. The image frame extraction module segments the cholecystectomy video to obtain image frames; the global feature extraction module inputs the image frame into a student Transform in a label-free self-supervised distillation architecture in the SMIL network architecture, and extracts global context features; the instance feature extraction module divides the image frame into a plurality of image blocks and inputs the image blocks into student Transformers in the multi-instance learning architecture to obtain an instance feature matrix; the local feature aggregation module inputs the instance feature matrix into an MIL Attention module to obtain local instance features; the feature fusion module performs feature fusion according to global and local features and dynamic weights to obtain specific fusion features corresponding to different standards; and the evaluation and prediction module inputs the fusion features into a linear classifier to obtain a CVS standard evaluation result. The system can be conveniently popularized to different hospitals and can adapt to operators of different styles or different devices.

The application discloses an interpretable image recognition method based on visual Transformer and prototype learning, and comprises the following steps: image preprocessing, normalizing and dividing a to-be-recognized image in training data into a plurality of non-overlapping image patches, then flattening each image patch into a one-dimensional vector; using different types of visual Transformer encoders to extract long-range dependence feature representation of the image patch; designing a prototype branch and composing a double-branch prototype layer and a single-branch prototype layer; according to the type of the visual Transformer encoder, calculating the similarity through the double-branch prototype layer or the single-branch prototype layer, obtaining an activation map through transformation, and processing the activation map; constructing a linear classifier to obtain an image recognition result; and interpreting the acquisition process of the image recognition result to complete the interpretable image recognition based on the visual Transformer and the prototype learning.

The invention relates to the technical field of automatic generation of laser radar point cloud training samples, in particular to an FPGA-based laser radar point cloud online sample self-generation method, which comprises the following steps of: performing Euclidean distance segmentation on input point cloud data based on an FPGA pipeline design, generating an initial segmentation point set by iteratively combining minimum distance point sets, and generating an initial segmentation point set; dynamically merging the initial segmentation point sets through an online equivalence strategy according to geometric features of the initial segmentation point sets; and carrying out initialization marking on the merged point set based on weak semantic knowledge, and filtering error samples by utilizing Bayesian cross validation and a linear SVM (Support Vector Machine) classifier. According to the method, full-process automation of point cloud segmentation, merging, marking and error sample filtering is realized, the problems that traditional sample generation depends on labeling and error samples cannot be filtered are solved, and the processing efficiency and the data quality are effectively improved.

The invention relates to a time-space diagram neural network autism classification method based on dynamic function connection and dynamic effective connection feature fusion, which can realize accurate classification of autism by using resting state functional magnetic resonance imaging data. The method comprises the following steps: firstly, constructing a dynamic function connection matrix and a dynamic effective connection matrix to respectively extract brain network diagram characteristics, and under the guidance of a dynamic effective connection network, forming fused brain network space-time connection characteristics by adopting a space-time fusion position Transform based on a cross attention mechanism; introducing a multi-layer perceptron to extract high-order image features in the brain network, embedding the high-order image features as node representation of a population graph, constructing edges of the population graph by using demographic information, realizing fusion of the magnetic resonance image features and the demographic information, and finally learning node embedding through a graph convolutional network to obtain the demographic information of the population graph. And autism classification is realized based on a linear classifier. Experimental results show that the provided method has excellent performance in autism diagnosis tasks, and the accuracy and robustness of diagnosis are remarkably improved.

Methods for obtaining black-box representations of machine learning models are disclosed when information about the models' internal states or parameters is inaccessible. By using the model's outputs instead of its internal states, the black-box representation is model-agnostic and provides a reliable and robust representation of the model through an external lens. The black-box representation is generated using responses from the model to a series of initialization and information-gathering questions, quantifying the model's confidence in the responses it has just returned. The black-box representation is then used as a training dataset for a linear classifier to learn performance metrics about the model.

The present invention relates to an adaptive training method for a P300 brain-computer interface classifier. Whenever a P300 brain-computer interface trial occurs, the brain-computer interface platform acquires EEG data. e t , and then determine whether this trial contains P300, and use EEG data e t and discriminative feedback y t Update the weight coefficients of the linear combination and the discriminant matrix of the current user's linear classifier. This method is conducive to dynamically updating and optimizing the P300 brain-computer interface classifier during operation.

The application discloses an industrial information physical system cross-layer anomaly detection method based on information physical feature fusion, acquires a data packet sequence of each communication event; for each data packet sequence, extracts coarse-grained features of a sequence level thereof, analyzes all data packets to obtain fine-grained features of a data packet level, and then fuses the coarse-grained features and the fine-grained features to obtain an overall feature vector of the sequence; taking the overall feature vector of the data packet sequence as observation data, projecting the observation data to a high-dimensional feature space by using a dictionary, learning the dictionary and a linear classifier; in online testing, obtaining an overall feature vector of a data packet sequence corresponding to a current communication event, obtaining sparse coding based on the dictionary, and outputting a current state label of the industrial system by using the linear classifier. The application improves real-time performance and accuracy of anomaly detection by effectively fusing information physical heterogeneous data.

The invention discloses a multi-granularity cascade fusion password attribution classification method and system. The method comprises the following steps: word segmentation of a password text: performing multi-granularity segmentation and feature representation on the password text based on a plurality of word segmentation methods; initialized coding: carrying out initialized coding on a password text word segmentation result based on a double-layer bidirectional long-short-term memory network; local feature coding: carrying out local feature coding based on a self-attention network and an initialized coding result to obtain local feature representations of various word segmentation representations; cascading fusion: fusing the local feature representations of the plurality of word segmentation representations into the unified feature representation; and classification: inputting the unified feature representation into a linear classifier so as to obtain a corresponding password text classification result. According to the ultra-short password text classification method, multi-dimensional fine-grained segmentation and representation are carried out on the ultra-short password text by utilizing multiple word segmentation methods, and the limited password text character features are fully utilized after the features of multiple representations are ingeniously fused, so that the purpose of efficiently classifying the ultra-short password text is achieved.