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

10results about How to "Save labeling time" patented technology

Multi-source pavement disease identification method based on language and image large model

The invention relates to a multisource pavement disease identification method based on a language and an image large model, and the method comprises the following steps: collecting related standards of road pavement disease detection, building a knowledge corpus, and reading a road disease image through a large language model LLM; when it is determined by the LLM that the crack crosses the whole image, cutting is carried out in the direction perpendicular to the crack by passing through the center point of the image, and the image is cut into two images; afterwards, all images and corresponding text prompts are input, target detection is carried out on pavement diseases by using a visual language large model VLM, and a prediction box is output; and performing fine adjustment on the SAM by using the image marked with the disease mask to obtain the RoadSAM adapted to the road disease segmentation task. According to the method, a large language model, a visual language large model and a segmentation large model are combined, corresponding fine adjustment is carried out on the segmentation large model SAM, dependence on a large amount of annotation data is reduced, and meanwhile unified segmentation of multi-source and multi-mode pavement disease images is achieved.
Owner:HEBEI UNIV OF TECH

Low-cost cell segmentation method, system and device based on active learning

The invention discloses a low-cost cell segmentation method, system and device based on active learning. The method comprises the following steps: selecting an initial cell image for preprocessing and bounding box labeling; forming a training set and constructing a target detection model for training so as to predict a bounding box of pictures of the verification set; taking the bounding box as a prompt for segmenting a cut model (SAM), and outputting a corresponding cell mask; adopting a Monte Carlo discarding method to predict the uncertainty of the unlabeled pictures, and selecting the pictures with higher uncertainty to be labeled by experts; adding the marked image data into the training set of the last round for updating, and repeating the training and sample selection of the model; and outputting a model weight and an index record after the required performance is achieved. According to the method, the data annotation cost of a cell segmentation algorithm is greatly reduced on the premise of ensuring the model performance.
Owner:ZHEJIANG UNIV

Remote sensing image photovoltaic identification method based on SAM model

The invention provides a remote sensing image photovoltaic identification method based on an SAM model, and relates to the technical field of computer vision, and the method comprises the steps: collecting and preprocessing a multi-resolution photovoltaic segmentation data set, and enabling the data set to be used for adjusting an image encoder; constructing an SAM segmentation network comprising the image encoder, the prompt encoder and the mask decoder, and inputting interactive click information and the data set into the SAM segmentation network; adding a segmentation output layer into the mask decoder to generate a preliminary segmentation result; and feeding the preliminary segmentation result back to an interaction module for interactive adjustment, and decoding again after adjustment to generate a final segmentation result.
Owner:QINGDAO INST OF BIOENERGY & BIOPROCESS TECH CHINESE ACADEMY OF SCI

Automatic labeling method, device and equipment and storage medium

The invention relates to the technical field of artificial intelligence and the technical field of data annotation, and discloses an automatic annotation method, device and equipment and a storage medium, and the method comprises the steps: obtaining a current object image, and obtaining a candidate region outputted by a candidate box generation network based on the current object image; performing feature extraction on the region image corresponding to the candidate region to obtain a current object feature; generating category similarity corresponding to each preset object category; selecting a preset object category with the maximum category similarity as a current object category corresponding to the current object feature; the method comprises the steps of obtaining a polygonal envelope corresponding to an object in a current object image, combining information of the polygonal envelope with a current object category, generating annotation information of the current object image, adding the annotation information to the current object image, and generating an annotated current object image. According to the invention, the marking efficiency of the current object image is improved.
Owner:浙江人形机器人创新中心有限公司

General apparent defect detection method based on visual large model

The invention relates to a general appearance defect detection method based on a visual large model in the field of industrial computer vision, and the method comprises the following steps: collecting and preprocessing a picture with defects, obtaining a defect picture feature T1, obtaining a target region feature # imgabs0 # and a target embedding feature T2 through coding calculation, storing # imgabs1 # and T2 in a memory bank, and obtaining a target embedded feature T1; a defect detection end-to-end neural network model is constructed and deployed, a to-be-detected picture is preprocessed and input into the defect detection end-to-end neural network model to obtain a detection result, and through combination of defect detection and a visual large model and dependence on a complex data set pre-trained by the large model, the training cost is reduced, and the detection precision is improved. Specific defect pictures are used as neural network prompts, the generalization of the method and the flexibility of model deployment are enhanced, the method adapts to different defect detection items, and customer requirements are met.
Owner:SHENSHI OPTOMETRY (SHANGHAI) TECHNOLOGY CO LTD

Method for generating three-dimensional space semantic occupancy grids at low cost

The invention relates to a method for generating a three-dimensional space semantic occupation grid at low cost. The method comprises the following steps: acquiring point cloud data and corresponding obstacle labeling information; according to the marked position of the obstacle, removing the marked point cloud of the obstacle in the point cloud data to obtain background point cloud; segmenting the generated background point cloud by using a ground segmentation algorithm, wherein all points in the background point cloud are divided into ground points and non-ground points; re-integrating the ground point cloud, the non-ground point cloud and the obstacle marking point cloud into a complete point cloud according to the semantics of the respective parts; a occupancy grid is generated for the generated semantic segmentation point cloud using an occupancy grid generation algorithm. According to the method, generation only needs to be carried out through an algorithm on the basis of point cloud obstacle annotation, and compared with point cloud semantic annotation, the cost is reduced by about 10 times on the existing annotation market at present. The technical scheme is simple in design, easy to implement and operate, and beneficial to being popularized and applied to actual scenes.
Owner:安徽中科星驰自动驾驶技术有限公司

Method and equipment for detecting surface water changes in radar remote sensing images based on deep learning

The present invention discloses a method and device for detecting surface water changes based on radar remote sensing images based on deep learning. A water body detection model is obtained by training a deep learning model based on water body annotation samples obtained by processing radar remote sensing images and cloud-free multispectral remote sensing images. The model can be used to obtain a binary raster image with the same size and resolution as the radar remote sensing image, and the previous and subsequent time-series binary raster images are compared at the pixel level to obtain a water body change detection result. The present invention overcomes the limitations of traditional surface water body detection methods based on optical remote sensing images, which have high requirements on weather conditions and need to collect multispectral remote sensing images without cloud cover.
Owner:EMERGENCY MANAGEMENT DEPT COMM & INFORMATION CENT

Logistics packaging recycling management method, device, equipment and readable storage medium

The present invention provides a logistics package recycling management method, device, equipment and readable storage medium, the method comprising: acquiring an image of each logistics package within a preset time period; extracting characteristic information on the logistics package according to the image of the logistics package to obtain characteristic information; dividing all the logistics package images into multiple groups according to the characteristic information, adjusting the logistics package images in each group so that the number of logistics package images in each group is consistent, obtaining a grouping result, and classifying and recycling all the logistics packages according to the grouping result. The method of the present invention can complete the grouping of all the logistics packages according to the characteristics of each logistics package, which is conducive to the subsequent recycling of the logistics packages.
Owner:HUNAN DAMEI LOGISTICS EQUIP MFG CO LTD

Processing method, device and equipment for sentence vector generation model based on artificial intelligence

The present invention discloses a processing method for a sentence vector generation model based on artificial intelligence, which is applied to the field of artificial intelligence technology and is used to solve the technical problems of low learning efficiency of text sentence vectors and lack of representation accuracy of sentence vectors. The method provided by the present invention includes: obtaining a text sample without a sample label; inputting the text sample into a BERT module in a sentence vector generation model to be trained, and outputting intermediate features of the text sample through the BERT module; selecting different discarding masks to perform dropout discarding operations on the intermediate features of the same text sample, and obtaining intermediate feature vectors corresponding to the same text sample and each discarding mask; determining the intermediate feature vectors belonging to the same text sample as positive samples, and determining the intermediate feature vectors belonging to different text samples as negative samples; training the sentence vector generation model according to the positive sample, the negative sample and the loss function, and obtaining a trained sentence vector generation model.
Owner:CHINA PING AN LIFE INSURANCE CO LTD

Method, apparatus, and computer-readable storage medium for labeling adjacent key mis-touch information

ActiveCN116402135BRealize automatic labelingSave labeling timeMachine learningInput/output processes for data processingInformation typeEngineering
This application belongs to the field of intelligent input technology, specifically relating to a method, apparatus, and computer-readable storage medium for labeling adjacent key mis-touch information. The method includes: acquiring log information generated by a user inputting information via a keyboard; parsing the log information to obtain target information corresponding to the same input cycle, wherein the target information includes typed information and on-screen information; acquiring the information type corresponding to the typed information and determining target rules based on the information type; performing adjacent key mis-touch analysis on the target information according to the target rules, and labeling adjacent key mis-touch information based on the analysis results, so as to train an adjacent key mis-touch recognition model based on the labeled adjacent key mis-touch information. This application can achieve automated labeling of adjacent key mis-touch information, shortening the information labeling time, reducing mis-labeling caused by manual labeling, accelerating the iteration speed of the adjacent key mis-touch recognition model, and improving the model's accuracy and input accuracy.
Owner:TENCENT TECHNOLOGY (SHENZHEN) CO LTD