12results about How to "Reduce labeling workload" patented technology

The application discloses a kind of blood vessels and lesion multi-task segmentation method based on ultra-wide-angle fundus image, belong to image processing field.The steps include as follows:S1.UWF fundus image dataset is obtained, and is assigned as training set and test set;S2.build multi-task semi-supervised learning network based on weight control mechanism, including encoder, decoder, cross-level non-local graph module and loss weight control mechanism, and training set image enters encoder and carries out feature extraction;S3.feature that encoder output is decoded and feature reconstruction by decoder part;S4.combined loss function and total loss are optimized to network by training, obtain the blood vessels and lesion multi-task segmentation model based on ultra-wide-angle fundus image;S5.data in test set are input into model, and segmentation result is obtained.The application can improve the ability of model to extract image detail features, realize fine and accurate multi-task segmentation.

The invention discloses a video data labeling method and device, electronic equipment and a storage medium. The method comprises the following steps: determining target video data and target information in the motion process of the intelligent equipment, wherein the target video data is video data collected by the intelligent equipment; determining a target image based on the target video data, and generating a target labeling result of the target image; determining a plurality of reference images based on the target video data, and generating respective target labeling results of the plurality of reference images based on the target labeling results of the target images; based on the target information, the target annotation result of the target image and the target annotation results of the multiple reference images, determining the confidence degree of the target annotation results of the multiple reference images, and based on the confidence degree of the target annotation results of the multiple reference images, determining the annotation result of the target object in the target video data. According to the scheme, the efficiency and accuracy of labeling the video data collected by the intelligent device with the body can be improved.

The embodiment of the invention provides a cross-platform graphical user interface (GUI) agent training data generation method and device, and privacy data of a user side does not need to be used. And inputting first GUI training data of the first platform into the large model to obtain a first sub-instruction sequence suitable for being executed in the second platform, and controlling a second controlled device of the second platform to execute an operation corresponding to the first sub-instruction sequence through the GUI intelligent agent. The first GUI training data comprises a first user instruction and a first action sequence, and is acquired when the controlled equipment of the first platform executes the first user instruction. And under the control of the GUI intelligent agent, the second controlled equipment collects corresponding operation data when executing corresponding operations of a plurality of time steps. When the second controlled equipment executes the corresponding operation of the multiple time steps, second GUI training data are generated according to the first user instruction and a second action sequence containing the operation data of the multiple time steps, and the second GUI training data are used as training data of a GUI intelligent agent of a second platform.

The present disclosure relates to a vehicle control method, device, computer equipment, storage medium and computer program product. The method comprises: obtaining to-be-predicted vehicle data of a vehicle; inputting the to-be-predicted vehicle data into a target perception model, and outputting a perception prediction result through the target perception model, wherein the target perception model is obtained after an initial perception model is optimized based on labeled vehicle data samples and is verified to pass, the labeled vehicle data samples include vehicle data samples labeled with a perception result label, the vehicle data samples are obtained when an initial perception result of the initial perception model has an error, and the perception result label has a correlation with a problem type corresponding to the error; and controlling the vehicle in a control mode matched with the perception prediction result. The present method can improve the accuracy of vehicle control.

The embodiment of the invention provides a student behavior data set labeling method, and the method comprises the steps: carrying out the seat region detection of a first image of a target classroom through employing a trained target detection model, and obtaining a seat detection frame set, and the first image comprises seats disposed in the target classroom; based on the shooting view angle of the first image, each seat detection frame in the seat detection frame set is allocated to a grid position formed based on a preset row number R and a preset column number C to generate a structured seat template, and R and C are integers greater than or equal to 1; a second image of the target classroom is matched with the structured seat template, student behaviors corresponding to each seat are identified, and the second image comprises students in the target classroom; and identifying the second image based on the behavior of the student corresponding to each seat, and adding the identified second image to the data set of student behavior identification.

The application discloses a method and system for realizing sleep staging by using Bayesian uncertainty, comprising: obtaining electroencephalogram signal data and preprocessing to obtain an unlabeled data set; using the data set to preliminarily train a Bayesian sleep network comprising two Bayesian sleep sub-networks with the same structure, outputting Bayesian uncertainty quantification results according to Bayesian entropy; selecting part of important samples according to the quantification results to label sleep staging labels, which are used for fine-tuning the network; passing the remaining samples which are not labeled through the two fine-tuned Bayesian sleep sub-networks respectively, outputting predicted sleep staging results, calculating Bayesian uncertainty quantification results, assigning weights to the predicted sleep staging results according to the quantification results and fusing the results to obtain overall predicted sleep staging results, selecting part of samples to correct the results and outputting final sleep staging results. The application can effectively reduce the dependence on sleep staging label labeling and improve sleep staging accuracy.

The application discloses a hyperspectral imaging data processing method for analyzing rock-soil particle shape and mineral components, and comprises the following steps: acquiring hyperspectral data of a mineral particle to be measured, and preprocessing the hyperspectral data of the mineral particle to be measured; obtaining a mineral component test set based on the preprocessed hyperspectral data of the mineral particle to be measured; calling a generated mineral component model weight file to identify and classify the mineral component test set, obtaining corresponding end-member mineral components, and statistically obtaining mineral components and relative contents of each mineral component based on the end-member mineral components, so as to realize mineral component analysis of the mineral particle to be measured. The application can accurately realize joint analysis of the shape and the mineral components of the mineral particle.

This invention relates to the field of microplastic identification technology, specifically to a method for identifying and classifying microplastics in water samples based on deep learning. The method includes the following steps: S1, image data acquisition and adaptive preprocessing; S2, determining whether the current process is in the training phase; if so, proceeding to steps S3-S6; otherwise, directly proceeding to step S6; S3, semi-automatic dataset construction based on saliency detection; S4, constructing a lightweight microplastic detection network; S5, model training and optimization: using transfer learning to load pre-trained weights, combined with online HSV domain enhancement and loss function optimization, to train the detection model and obtain a high-precision microplastic detection model; S6, inference and post-processing screening. This invention improves the accuracy and stability of identifying microplastics of different sizes and shapes through adaptive preprocessing, semi-automatic annotation, and lightweight model design.

This invention discloses a method and system for few-sample instance segmentation and part detection in small-batch scenarios. The method includes: determining the detection area; acquiring a video of the object to be labeled and a reference background image from a fixed camera position, and assigning labels to the object to be labeled; discretizing the object video into multiple images, inputting them along with the reference image into a reference-based foreground segmentation model to obtain the Alpha mask of the target foreground, and generating instance labels by combining the sample labels; then forming a part training set using conventional instance segmentation data augmentation techniques; training a lightweight improved Mask R-CNN model using the part training set and a human key part training set to obtain the model weights for this batch; designing a human-machine collaborative detection process, performing frame extraction processing on the video input, and simultaneously detecting human key parts and parts from a fixed viewpoint to obtain instance information of the parts.

This invention relates to the field of power grid information security technology, specifically a method, system, device, and storage medium for detecting malicious software in the power grid Internet of Things (IoT) based on active learning. The method involves acquiring application samples from power grid IoT nodes, extracting static features, dynamic features, and power grid context information to form a multi-dimensional feature vector, and organizing these into data blocks according to timestamps. Classification uncertainty scores are calculated from an unlabeled sample pool, a detector committee is constructed to calculate consensus entropy, and the sample with the most information content is selected by combining the two scores and submitted for expert annotation. A random forest classifier is trained to build a detection model. The F1 score of the current data block is evaluated; annotation stops when a preset threshold is reached and is applied to the next data block. Model performance changes are monitored, and when performance degradation is detected, batch retraining, rolling back historical configurations, or incremental updates are performed based on the evolution of the threat environment. The method reduces annotation costs through sample selection and addresses the conceptual drift problem of the power grid threat environment through an adaptive update strategy.

The invention discloses a traditional Chinese medicine classical knowledge graph construction method based on dynamic semantic analysis, and the method comprises the steps: carrying out the semantic analysis processing of an input traditional Chinese medicine classical book according to a pre-constructed dynamic semantic library, and obtaining a standardized text; wherein the dynamic semantic library represents a database of standard semantics of traditional Chinese medicine terms, knowledge extraction processing is carried out on the standardized text according to preset term matching sentence patterns to obtain a structured result, and the term matching sentence patterns comprise high-frequency sentence patterns and term matching rules appearing in traditional Chinese medicine classics; and constructing a knowledge graph according to the accessed time sequence data and the structured result. According to the invention, the accuracy of traditional Chinese medicine term and metaphor identification can be improved, and the precision of semantic understanding is improved.

This invention relates to a weakly supervised image target localization method based on multi-scale salient feature fusion, belonging to the field of computer vision. To address the problems of cumbersome ROI annotation and insufficient CAM activation in small target images, this invention focuses on optimizing the output class activation map of a classification network under weak supervision. This invention relates to information fusion at two levels: ① Since the feature maps at the lowest level of a convolutional neural network have weak semantic information but strong positional information, they can be fused with the feature maps at the highest level to obtain the final feature map of the classification network; ② Because the classification network has different sensitivities to ROIs at different scales, the resulting class activation maps also differ. Therefore, fusing complementary object information from different activation maps can improve the localization of target regions in the image, thereby generating more accurate pseudo-labels for segmentation tasks.