14 results about "Annotation database" patented technology

The invention provides an automatic action trajectory labeling system and method based on a visual language model, and relates to the field of visual technical models. The automatic action trajectory labeling system based on the visual language model comprises a key frame extraction module, a VLM semantic labeling module, a historical labeling database, a diffusion model optimization module and an reflection correction module, and the key frame extraction module extracts a key frame sequence from an input video through optical flow analysis and a TSN network scene change detection algorithm; and the VLM semantic annotation module is used for outputting a candidate segmentation point set containing action starting and ending points, the VLM semantic annotation module adopts a multi-modal VLM, takes a key frame image and a context text instruction as input, and generates a preliminary semantic tag and a corresponding vision-language embedding vector. According to the method, multiple bottlenecks of the existing action track labeling technology are solved, the labeling precision is remarkably improved, and the target of time boundary optimization is achieved.

The application relates to a mycobacterium tuberculosis drug resistance phenotype prediction method and device and a computer device. The method comprises the following steps: determining a to-be-predicted mutation site which does not exist in a mutation phenotype annotation database from at least one mutation site of a target gene fragment of to-be-processed mycobacterium tuberculosis according to gene sequencing data of the mycobacterium tuberculosis; converting protein structure change data of the to-be-predicted mutation site according to protein structure data corresponding to the to-be-predicted mutation site; inputting the protein structure change data of the to-be-predicted mutation site into a drug resistance phenotype prediction model to obtain a predicted drug resistance phenotype of the to-be-predicted mutation site; and determining a drug resistance phenotype prediction result of the to-be-processed mycobacterium tuberculosis based on the predicted drug resistance phenotype of the to-be-predicted mutation site and drug resistance phenotypes of the mutation sites except the to-be-predicted mutation site. The method can effectively improve the mycobacterium tuberculosis drug resistance phenotype prediction efficiency.

Creating a medical report, indicative of a longitudinal examination includes receiving an input longitudinal report data structure having longitudinal data elements, which are associated with annotation elements of a medical annotation database that is indicative of relationships between medical concepts. A set of reference report data structures is received, which has reference data elements associated with annotation elements of the medical annotation database. For each longitudinal data elements, it is determined whether the longitudinal data element can be linked with a reference data element of at least one reference report data structure from the set of reference report data structures. The determining is made based on the associated annotation elements and based on the medical annotation database. An interface element indicative of a pointer from the input longitudinal report data structure to the reference report data structure having the linked reference data element, is outputted responsive to the determining.

The invention relates to the technical field of cultural relic identification, in particular to a porcelain identification system and method based on artificial intelligence, and the system comprises a short-wave infrared camera module which is used for collecting an infrared transmission image of porcelain in a wave band of 760-1mm; the visible light shooting module is used for acquiring a porcelain surface visible light image; the image synchronization control module ensures that the short infrared shooting module and the visible light shooting module synchronously take images at the same angle and position; and the AI image processing module is used for analyzing the glaze layer reflection characteristic, the carcass material reaction characteristic and the aging structure difference of the double-spectrum image based on a deep learning architecture, comparing with the annotation database, and outputting the kiln mouth, the age and the authenticity judgment result of the porcelain. According to the porcelain recognition system and method based on artificial intelligence, the aging characteristics, the kiln mouth and the age of porcelain are presumed through AI image analysis, and the interpretation efficiency and accuracy of the porcelain are improved.

The invention provides a robot body data labeling method and device based on an artificial intelligence large model, and the method comprises the steps: obtaining the multi-modal original data of a robot during task execution, and carrying out the standardization processing of the original data, and obtaining the standardized data; performing multi-modal semantic understanding and analysis on the standardized data based on an artificial intelligence large model, and reversely generating a structured task description; and generating a structured labeling result according to the structured task description, and storing the structured labeling result into a labeling database. According to the method and the device provided by the invention, data format conversion and data enhancement are performed on the robot body data through the large model, so that the consistency of the data annotation source data file and the annotation result is realized. A unified semantic understanding standard is provided, subjective deviation of manual annotation is eliminated, semantic and time sequence synchronization of language, vision and action is guaranteed through an end-to-end multi-mode alignment mechanism, the labor cost is greatly reduced through an automatic process, and large-scale data production is supported.

The invention discloses a feature labeling method and device based on vision, a medium and a product, and relates to the field of image processing. The method comprises the following steps: calculating the confidence of a target feature of a to-be-detected image through a feature recognition model; if the confidence coefficient is smaller than a preset confidence coefficient threshold value, determining that the to-be-detected image is a to-be-calibrated image; extracting a first preset number of first feature sequences from a historical annotation database; acquiring an annotation instruction of a quality inspector on the to-be-calibrated image; if the labeling instruction shows that the target feature is the newly added feature of the second feature sequence, labeling the to-be-detected image as a defect type corresponding to the second feature sequence, and adding the to-be-detected image to the tail end of the second feature sequence to obtain a third feature sequence; and incremental training is carried out on the feature recognition model to obtain an updated feature recognition model, and the updated feature recognition model is used for image recognition and annotation. By implementing the technical scheme provided by the invention, the robustness of feature labeling can be improved.

The invention discloses a radiotherapy scheme decision-making method and system based on a tumor classification labeling database. The method comprises the following steps: collecting historical spinal tumor radiotherapy data of a patient; labeling the historical spinal tumor radiotherapy data to obtain a classification labeling database; historical text data, historical medical image data and historical multi-source time sequence data of multiple systems in a hospital are collected, and multi-modal fusion data are obtained; mapping the multi-modal fusion data and the classification annotation database to obtain historical diagnosis and treatment data; collecting current case data; and matching the current case data with the historical diagnosis and treatment data to obtain a radiotherapy scheme decision result. The scheme of the invention effectively solves the problems of insufficient utilization of existing spinal tumor radiotherapy data and lack of data support in decision making.

The present disclosure relates to a surgical navigation system (1) having an annotation function for inputting position-related, particularly visual, medical annotations or records and / or for outputting medical annotations during a surgical procedure on a patient (P), the surgical navigation system (1) comprising a display device (2), particularly an operating monitor, for visual output of medical annotations, at least one robot (4) as a positioning system, particularly a multi-link robotic arm (8) movably connected to a robot base (6), and a robot head (10) arranged at an end of the robotic arm (8), a stereoscopic recording system (12), particularly a stereo camera, particularly preferably a stereomicroscope, arranged on or within the robot head (10), the stereoscopic recording system comprising at least two recording units (14) arranged spaced apart from each other for performing a 3D recording (3DA) of a structure or surface for the recording system (12) and providing it in a computer-readable format together with recording parameters, a tracking system (16) adapted to detect and track the orientation of the recording system (12), particularly indirectly via the orientation of the robot head (19), and provide the orientation of the recording system (12) in a computer-readable format, an annotation system (18) comprising a storage device in which a central digital annotation database having a global coordinate system for annotation objects is stored, and a control unit (20) adapted to process the provided 3D recording (3DA) together with the recording parameters and the provided orientation, determine the orientation of the 3D recording (3DA) in the global coordinate system based on the provided recording parameters and the orientation of the recording system (12), and output at least one annotation object of the annotation database in a superimposed display (U) via the display device (2) at least at a correct position, particularly in a correct orientation, within the 3D recording (3DA) to improve navigation. The present disclosure relates to a navigation method and a computer-readable storage medium according to the independent and dependent claims.

The invention discloses a computer-aided antibody off-target effect prediction method and system. The method mainly comprises the following steps: constructing an antibody-antigen structure and function annotation database, carrying out holoproteome structure search on an antibody by adopting a 3Di / aa space structure descriptor, and carrying out multi-dimensional comprehensive evaluation on a search result in combination with structural similarity. According to the method, high-flux and low-cost prediction of the antibody off-target effect can be realized, the limitation that the traditional method only depends on sequence alignment or single structure scoring is broken through, the prediction accuracy is improved, the risk caused by off-target is reduced, and the research and development period of antibody drugs is remarkably shortened.

The application provides a method and related equipment for screening potential pathogenic variants and genes, the method comprising: obtaining a gene sequencing variant file of a subject, the gene sequencing variant file containing variant site information; selecting a first annotation database to annotate the variant site information in the gene sequencing variant file to obtain a variant annotation file; screening the variant site information in the variant annotation file according to a predetermined variant screening condition to obtain a variant screening file; and selecting a second annotation database to supplement the annotation of the variant site information in the variant screening file to obtain a potential pathogenic variant file. The method and related equipment for screening potential pathogenic variants and genes provided by the application have high screening efficiency, can filter out a large number of variant sites with less biological significance, retain more valuable variant site information, provide a good foundation for downstream analysis, reduce the workload of interpreters, and save time and effort.

The invention relates to the technical field of microbiology, in particular to a microbial metagenome annotation database suitable for a specific application scene and a construction method, device, system and application thereof. Compared with the prior art, the method for constructing the microbial metagenome annotation database aiming at the specific application scene is provided, the accurate annotation capability of the database on the seed level and the strain level of the specific application scene can be improved by adopting the construction method, and the method has the characteristics of being complete and precise.

The invention discloses an Escherichia coli alkali-resistant target spot and application thereof, an inhibition model of pH on Escherichia coli E.coli AG1 is established by setting LB culture media with different pH values, and it is found that the inhibition degree of cell growth is positively correlated with the rise of pH. Under the condition that the pH value is equal to 10, the escherichia coli overexpression library ASKA is subjected to high-throughput screening, and 43 alkali-tolerant elements are obtained after multiple rounds of screening. The obtained gene is subjected to KEGG / GO enrichment analysis through an online gene function annotation database DAVID, an EcoCyc database and related literatures are looked up, an alkali tolerance mechanism is predicted, and it is found that the alkali tolerance mechanism is possibly related to biological membrane formation, intracellular pH steady state maintenance, intracellular ROS level reduction and the like. Through overexpression of the genes, the strain can adapt to an environment under alkali stress, the alkali tolerance of the strain is improved, and the strain has a good application prospect.

The invention discloses an analysis method for automatically processing metagenome high-throughput sequencing data binning to obtain a viral genome, which comprises the following steps: acquiring second-generation sequencing original data, and performing data quality control to obtain Reads to be analyzed; a plurality of Reads are assembled into an overlapping group through overlapping of fragments, statistics is conducted on the overlapping group, and a length distribution diagram is drawn; reconstructing a viral genome based on contig binning, and generating a box after sequence clustering; constructing a three-dimensional matrix of all boxes according to the attributes of the sequence, and predicting and identifying the species category of each box; carrying out quality identification on the virus box bin, and detecting and rejecting virus host genes; counting the abundance of the virus box bin in each sample to generate a virus box bin abundance table; constructing a virus annotation database, completing species annotation, constructing a virus evolutionary tree, and finally generating a report. The invention further discloses an analysis system and application for implementing the analysis method.

The invention belongs to the technical field of artificial intelligence, and particularly relates to a genetic interpretation system device integrating multi-variation annotation and intelligent scoring, which comprises a variation information acquisition module used for collecting illness conditions of each member in a family to be analyzed and exon variation site information of each member, performing comparative analysis according to the collected information and data in a preset database to determine a variation type, and generating a genetic variation information data set; the annotation database construction module is used for constructing an annotation database and storing and managing a genetic variation information data set; the intelligent scoring module is used for calling the genetic variation information data set stored in the annotation database and generating an analysis result; and the report generation module is used for generating a report according to the analysis result. The system can assist doctors in genetic disease analysis, and a genetic diagnosis report is automatically generated through the report generation module, so that the credibility of the report is improved.