84 results about "Pathological" patented technology

The invention belongs to the technical field of noise detection, and provides a detection method for pathological voice. The method comprises the following steps of collecting the voice of a patient to be detected, conducting characteristic parameter extraction and selection on the collected voice signal, enabling optimized parameters to enter a constructed classifier model for conducting disorder grade evaluation, and outputting the detected voice disorder grading result. According to the detection method for the pathological voice, a computer and scientific judging standards are used, a professional voice processing algorithm is adopted, a doctor can be partially or completely replaced for diagnosing the patient, the result is used as the diagnosis reference for the doctor, and the contingency of the diagnostic process is reduced to the greatest extent. In addition, the detection method is easy to implement, convenient to use and high in diagnosis accuracy, an ordinary medical worker can master the detection method through simple training, the defect that medical resources are not enough in remote areas and small cities is overcome to some extent, and the disease of the patient can be diagnosed nearby and treated as soon as possible. Moreover, the specific and quantified grading mode is provided for the voice disorder, corresponding data logs are provided at each stage in the treatment process of the patient, the doctor can completely track and know the state of the disease through the data, and the treatment process of the patient is ensured to the greatest extent.

The invention relates to a pathological section interpretation method, which is suitable for a pathological section interpretation system. The method comprises the following steps that S10, a microscope camera collects a pathological image of a solid PD-L1 immunohistochemical staining section under a microscope, and sends the pathological image to a processor; S20, the processor preprocesses the pathological image, so that the HSV color space of the pathological image is consistent with a set threshold value; S30, the processor processes, detects and marks the preprocessed pathological image,and outputs a marked pathological image; S40, the display receives and displays the marked pathological image from the processor. The invention further discloses a pathological section interpretationsystem. A computer-readable storage medium stores a computer program for use in conjunction with a pathological section interpretation system. According to the application, the artificial intelligencemethod is used to improve the diagnosis efficiency of doctors and accurately and quickly provide PD-L1 positive scores to assist doctors in clinical judgment.

The invention relates to a pathological whole process quality control and information management system. The system comprises a pathological application module, a sample receiving module, a material sampling module, a slice production module, a diagnostic module, a quality evaluation module, a sample database management module, a wax block management module, a slice management module and the like.Through reasonably setting association modes between a system module and the modules, effective pathological department whole process quality control management and clear rights and responsibilitiesare realized, automatic circulation among inspection items in the pathology department is realized, needs of the pathological department for unified information storage and management and disaster recovery backup are satisfied, and remote access, remote maintenance and mobile login management are realized.

The invention provides methods, compositions, and systems for diagnosis, prognosis, evaluation of status, and/or determination of treatment for pathological conditions.

The invention relates to the technical field of medical treatment, and discloses a multi-modal medical data fusion evaluation method and device, equipment and a storage medium. The method comprises the steps of obtaining multi-modal to-be-evaluated medical data of a target object; performing feature extraction on the to-be-evaluated medical data of each mode to obtain a plurality of feature vectors, and performing fusion to obtain a fused feature vector; and inputting the fusion feature vector into a trained multi-modal fusion evaluation model to obtain an evaluation result output by the model. According to the method, feature extraction and feature fusion are carried out on the multi-modal medical data based on artificial intelligence to obtain the fusion feature vector, and the illness state remission degree of the target object is predicted and evaluated by using the multi-modal fusion evaluation model based on the fusion feature vector, so that the illness state remission degree under a pathological level can be assisted to be accurately evaluated, the judgment accuracy is improved, and medical risks are reduced. The invention further discloses evaluation of multi-modal medical data fusion.

The invention discloses a thyroid cell pathological section malignant region detection method based on deep learning. The method mainly comprises the following steps: performing pathological section on thyroid cells; carrying out digital processing on the image of the pathological section on a microscope, and smearing with different coloring agents to obtain a colored pathological section; cuttingthe complete pathological section into dices with proper sizes as the input of a deep neural network model; screening out invalid dices of the pathological sections; carrying out benign and malignantclassification on the pathological sections subjected to slicing and preliminary screening by adopting a weakly supervised learning method; constructing a random forest-based machine learning methodby utilizing a false positive removing scheme to remove false positive from a prediction result of benign and malignant classification; therefore, the detection accuracy can be further improved. A pathological section high-risk area display step includes normalizing the malignant prediction probability of each block and mapping the malignant prediction probability of each block into the original image to generate a thermodynamic diagram, and providing more intuitive visual display for pathologists.

A CNN is applied to a histological image to identify areas of interest. The CNN classifies pixels according to relevance classes including one or more classes indicating levels of interest and at least one class indicating lack of interest. The CNN is trained on a training data set including data which has recorded how pathologists have interacted with visualizations of histological images. In the trained CNN, the interest-based pixel classification is used to generate a segmentation mask that defines areas of interest. The mask can be used to indicate where in an image clinically relevant features may be located. Further, it can be used to guide variable data compression of the histological image. Moreover, it can be used to control loading of image data in either a client-server model or within a memory cache policy. Furthermore, a histological image of a tissue sample of a tissue type that has been treated with a test compound is image processed in order to detect areas where toxic reactions to the test compound may have occurred. An autoencoder is trained with a training data set comprising histological images of tissue samples which are of the given tissue type, but which have not been treated with the test compound. The trained autoencoder is applied to detect tissue areas by their deviation from the normal variation seen in that tissue type as learnt by the training process, and so build up a toxicity map of the image. The toxicity map can then be used to direct a toxicological pathologist to examine the areas identified by the autoencoder as lying outside the normal range of heterogeneity for the tissue type. This makes the pathologists review quicker and more reliable. The toxicity map can also be overlayed with the segmentation mask indicating areas of interest. When an area of interest and an area identified as lying outside the normal range of heterogeneity for the tissue type, and increased confidence score is applied to the overlapping area.

The invention provides a pathological number identification method and device, an information identification method and device and an information identification system and relates to the field of artificial intelligence. The method comprises steps that a pathological slide image containing a label is obtained, and the label comprises a pathological number; edge detection of all objects in the pathological slide image is performed, and an outline corresponding to the label is determined according to a detection result; the label is corrected according to the position information of the contourto obtain a target image corresponding to the label; a pathology number area in the target image is extracted, and the pathology number in the pathology number area is identified to obtain the pathology number. The method is advantaged in that information identification efficiency and accuracy of the information identification result can be improved, and the labor cost is further reduced.

The invention relates to the field of digital medical treatment, and provides a pathological data analysis method and a device, computer equipment and a storage medium. The method comprises steps: performing feature extraction on pathological data to obtain specified feature data related to the heart failure occurrence risk; obtaining structured feature data of a user related to the heart failureoccurrence risk; splicing the specified feature data and the structured feature data to obtain fused feature data; taking the fused feature data as input of a preset attention module, generating attention weights corresponding to the fused features in the fused feature data one by one, and performing weighted summation on the fused features in the fused feature data according to the attention weights to obtain an output result; and inputting the output result into a preset classification module, and performing normalization processing on the output result to generate a heart failure occurrenceprediction probability of the user. According to the invention, the processing efficiency and prediction accuracy of predicting the heart failure occurrence risk of the user are improved.

The invention relates to a thyroid frozen section diagnosis method, and the method is suitable for a thyroid frozen section diagnosis system. The method comprises the following steps that S10, a microscope camera acquires a pathological image of a thyroid frozen section under a microscope and sends the pathological image to a processor; S20, the processor preprocesses the pathological image, so that the HSV color space of the pathological image is consistent with a set threshold value; S30, the processor detects the preprocessed pathological image, determines the range of the tumor area, outputs the pathological image marked with the range of the tumor area, or judges the yin and yang of the tumor in the pathological image, and outputs the pathological image marked with the negative and positive of the tumor; S40, the display receives and displays the marked pathological image from the processor. According to the detection method provided by the invention, the expenditure of a high scanner is saved, the diagnosis time in thyroid surgery is shortened, the surgery time is shortened, and the convenience is brought to rapid and accurate treatment.

The invention discloses a disease risk analysis system based on a pathological mode and an attention mechanism, wherein the system comprises an information collection device and a server; the information collection device is in communication connection with the server; the information collection device transmits the acquired demographic characteristic data information of a plurality of patients and the acquired electronic medical record data information of a target disease to the server; a portrait construction module, a graph embedding algorithm module, a medical record information extractionmodule, a training set construction module, a decision tree integration model module, a vectorization module, an attention mechanism module, a representation vector calculation module, a splicing module and a logistic regression analysis module are arranged in the server, and the server obtains a result required by health early warning based on the modules. According to the invention, big data service support is provided for health early warning, and the system has the advantage of accurate analysis result.

One technical scheme of the invention provides a pathological image classification method based on a multi-scale domain adversarial network. Another technical scheme of the invention is to provide a pathological image classification system based on a multi-scale domain adversarial network, which is characterized by comprising a preprocessing module; a single-scale feature extraction module; an overall feature extraction module; a multi-scale attention module; and a model evaluation module. According to the method, on one hand, WSI multi-scale feature information is combined, on the other hand, the influence of different dyeing effects on a prediction result is inhibited by using the domain adversarial network, and the volatility of the pathological image caused by dyeing is reduced, so that a system for assisting the pathological doctor to classify the pathological image in a manner of simulating the actual operation process of the pathological doctor is provided.

The invention provides a random survival forest-based liver cancer postoperative recurrence prediction method and a storage medium. The method comprises the following steps: acquiring clinical data and recurrence time of each case, the preset grouping dimension comprising basic factors of the patient, preoperative examination factors and postoperative pathological factors; obtaining a data set according to the clinical data, wherein the data set is composed of preset grouping dimensions corresponding to each case; and according to the data set and the recurrence time of each case, a random survival forest algorithm is adopted to construct a corresponding liver cancer postoperative early recurrence prediction model. According to the method, the postoperative recurrence probability of liver cancer of an individual patient can be accurately predicted, and the postoperative attention can be better determined; active prevention is facilitated; particularly, for medical institutions, medical staff can be helped to accurately screen out high-risk relapse patients after the liver cancer operation, intervention in the early relapse stage is facilitated, and postoperative follow-up visit and treatment are guided.

The invention belongs to the field of partial multi-mark learning and data mining, and particularly relates to a patient screening and marking method based on partial multi-mark learning. The method comprises the following steps: acquiring pathological sample data of a patient, inputting the pathological sample data into a trained medical text semantic information big data prediction model based on partial multi-mark learning, predicting a disease type and a disease probability of the patient, and marking the patient according to the disease type and the disease probability of the patient. According to the method, the problem of classification class imbalance is further processed, a more accurate marking result can be predicted, a patient can perform health management according to the marking result, a doctor can perform next diagnosis on the patient according to the result, and the method has good social benefits and economic benefits.

The invention discloses a pathological picture-based renal clear cell carcinoma molecular feature prediction and prognosis technology. The renal clear cell carcinoma molecular feature prediction and prognosis technology comprises the following parts: 1, feature extraction of the pathological picture; 2, predicting the molecular characteristics of the patient through the pathological picture; 3, predicting the lifetime of the patient by integrating a single pathological picture and a pathological picture with multiple omics; according to the method, the pathological picture of the patient can be quickly and economically quantified, and the survival time of the patient with existing gene, transcription or proteomics data can be quickly, economically and more accurately judged when the important mutation state, molecular subtype attribution and survival of the patient are predicted.

A method of identifying a gene associated with a disease or pathological condition of the disease includes: a) obtaining a first group of exome sequences from a first population suffering from the disease or pathological condition and a second group of exome sequences from a second population not having the disease or pathological condition; b) identifying one or more variants in the first group by comparing it with the second group, and optionally with a public database, to generate a first set of variant data; c) applying a variant quality score calibration tool with a truth sensitivity threshold to remove false-positive variants having a sensitivity lower than the threshold and background variants from the first set of variant data so as to obtain a second set of variant data; d) removing synonymous variants from the second set of variant data to obtain a third set of variant data; and e) identifying one or more deleterious variants from the third set of variant data using a gene burden analysis, optionally generating a fourth set of variant data.

The invention provides a pathological knowledge graph construction method and device. A pathological data knowledge base is built according to medical text data of different corpus sources, and the pathological data knowledge base comprises disease class entity pairs and a first entity relationship used for representing the association between two disease entities in the disease class entity pair; based on the medical data, a medical rule knowledge base is created according to a preset medical rule, and the medical rule knowledge base comprises interpretation type entity pairs and a second entity relationship used for representing association between two interpretation entities in each interpretation type entity pair; equivalent class entities in the disease class entity pair and the interpretation class entity pair are determined; based on the equivalence class entities, the pathological data knowledge base and the medical rule knowledge base are fused into a target knowledge base; and a target pathological knowledge graph is formed based on the target knowledge base for storage.

The invention discloses a renal clear cell carcinoma molecular feature prediction and prognosis technology based on a pathological picture. The renal clear cell carcinoma molecular feature predictionand prognosis technology comprises the following steps: 1, carrying out feature extraction of the pathological picture; 2, predicting the molecular characteristics of the patient through the pathological picture; 3, integrating a single pathological picture and a pathological picture into multiple omics to predict the lifetime of the patient. According to the method, pathological pictures of patients can be quantified quickly and economically, important mutation states, molecular subtype affiliation and survival time of the patients can be predicted, and the survival time of the patients withexisting gene, transcription or proteomics data can be judged quickly, economically and more accurately.

The invention provides a method and system for identifying epilepsy focus of temporal lobe epilepsy caused by hippocampus sclerosis and/or predicting pathological typing of the epilepsy focus of temporal lobe epilepsy caused by hippocampus sclerosis, and the method for locating the epilepsy focus of temporal lobe epilepsy caused by hippocampus sclerosis comprises the following steps: acquiring analysis data, including acquiring PET/MR dynamic continuous brain imaging of < 11 > C-choline, < 18 > F-FDG and < 11 > C-FMZ in the attack interval of a patient to be identified; carrying out reconstruction of analysis data, including performing data reconstruction on PET/MR dynamic continuous brain imaging of 11C-choline, 18F-FDG and 11C-FMZ in the attack interval of the to-be-identified patient, and obtaining reconstruction data synchronized with the analysis data; inputting the analysis data and/or the reconstruction data to an epilepsy focus positioning model, and processing and analyzing the analysis data and/or the reconstruction data by the epilepsy focus positioning model to obtain an output image for indicating an epilepsy focus area; and outputting an output image for indicating the epilepsy focus area.

The embodiment of the invention discloses a pathology critical value early warning method based on a pathology knowledge graph. The method comprises the following steps: acquiring pathology text information from a pathology report; extracting a target entity from the pathological text information; performing matching analysis on the target entity by utilizing a preset pathology knowledge graph, and determining whether a pathology critical value exists in the pathology report or not; if the pathological critical value exists, early warning reminding is carried out, so that timely early warning of the pathological critical value is realized, and a pathologist and a clinician are timely reminded to pay attention preferentially and early and carry out clinical treatment of the next step, and clinical treatment is timely carried out on the patient to the greatest extent and the timely treatment rate of the patient is greatly improved, and therefore, the pathological critical value early warning efficiency and the medical quality are improved. In addition, the invention also provides a pathology critical value early warning system based on the pathology knowledge graph, computer equipment and a storage medium.

The invention provides a pathological characteristic probability distribution prediction method based on pathological data, which comprises the following steps: S1, acquiring pathological data of a patient, performing dimension reduction processing on the pathological data to obtain corresponding characteristic data, and pre-configuring a plurality of characteristic tags; s2, for each feature tag, predicting a prior probability corresponding to the feature tag according to the feature data and at least one statistical learning model, and then processing according to the prior probability to obtain a first posterior probability, predicting a second posterior probability corresponding to each feature tag according to each feature data and at least one machine learning model; and S3, processing according to the first posterior probability and the second posterior probability to obtain a probability mean value, and forming pathological feature probability distribution of the patient by all the feature tags and the corresponding probability mean value. The method has the advantages that the pathological feature probability distribution of the patient is predicted according to the pathological data to assist doctors in classifying the pathological data of the patient, and the diagnosis efficiency is improved.