37 results about "Nodular thyroid" patented technology

The invention relates to a medical image processing technology, and aims to provide an ultrasonic thyroid nodule benign and malignant feature visualization method based on deep learning. The method comprises the following steps: collecting case data with both a thyroid nodule ultrasonic image and a clinical operation pathological result, distinguishing benign and malignant conditions, and markinga nodule region to generate a mask image; selecting a basic structure of a deep convolutional neural network, and performing segmentation pre-training on the mask image data of all thyroid nodules; initializing a basic network by using the model parameters, and constructing a deep convolutional neural network for identification; training and verifying in a folding intersection mode to obtain a benign and malignant recognition model; and inputting a test image, predicting an identification result by using the benign and malignant identification model, and generating a malignant feature visualization image. According to the invention, the relation between the benign and malignant probability of the nodule and the image area can be visually observed. A user can better analyze the image characteristics of the ultrasonic thyroid nodule, clinical puncture examination is further guided, and the success rate of a puncture operation is increased.

The invention discloses a thyroid nodule automatic detection model construction method, system and device based on a convolutional neural network, and the thyroid nodule automatic detection model construction method comprises the steps: carrying out the noise reduction of thyroid ultrasonic image data, and obtaining a thyroid ultrasonic image training data set; based on the training data set, using a Yolov3 network to train a thyroid nodule detection model; based on the training data set, using a Resnet network to train a thyroid nodule benign and malignant identification model; and fusing thethyroid nodule detection model and the thyroid nodule benign and bad recognition model to generate a thyroid nodule automatic detection model.

The invention provides a thyroid nodule analysis system based on elastic ultrasonic imaging, and relates to the technical field of computer-aided analysis. The system comprises: a data obtaining module which carries out thyroid nodule selection on an obtained elastic ultrasonic image, and obtains a thyroid nodule image; an edge recognition module used for carrying out edge recognition on the thyroid nodule image to obtain a nodule edge image; a feature analysis module used for respectively carrying out feature analysis on the thyroid nodule image and the nodule edge image to obtain a pluralityof image feature parameters; a feature selection module used for respectively calculating inter-class distances of the image feature parameters and adding the inter-class distances into a one-class interval sequence; and a nodule analysis module used for extracting the image feature parameters of the preset number of inter-class distances sorted in the front, and training to obtain a thyroid nodule state recognition model for subsequent thyroid nodule state recognition by taking the image characteristic parameters as input and the nodule state as output. The thyroid nodule analysis system hasthe advantage of effectively improving thyroid nodule state recognition accuracy.

The invention discloses a thyroid nodule semi-supervised segmentation method based on an attention mechanism. The method comprises the following steps of: 1, carrying out preprocessing of a thyroid ultrasonic image, and removing an edge information region in the image; 2, constructing a semi-supervised segmentation neural network, performing classification and segmentation prediction tasks on theultrasonic image, and adjusting a network structure to adapt to a specific application scene; 3, adding an attention mechanism into the semi-supervised segmentation neural network to improve the network effect; 4, measuring the performances of a semi-supervised segmentation algorithm and an existing full-supervised segmentation algorithm in the field of thyroid nodule auxiliary diagnosis through an intersection-parallel ratio and a Dice coefficient; and 5, continuously reducing the number of the pixel-level labels, and observing the change condition of the network performance. According to theinvention, the thyroid nodule semi-supervised segmentation method based on an attention mechanism benefits from the semi-supervised effect of a small number of pixel-level labels while keeping the high segmentation performance of the semi-supervised segmentation model, learns the real benign and malignant characteristics of the nodules and improves the benign and malignant classification capacity.

The invention provides a method for detecting nodules in a thyroid ultrasound image based on deep learning. The method comprises the steps of preprocessing the thyroid ultrasound image; extracting features of the preprocessed thyroid ultrasound image to obtain a feature image; respectively inputting the obtained feature images into corresponding classification and regression structures, and obtaining specific position information of a thyroid nodule region in each feature image; for the classification loss, the central point distance regression loss and the offset loss generated by calculation of the feature images input into the corresponding classification and regression structures, obtaining the total loss of the to-be-trained model through weighted summation calculation; and training and testing the to-be-trained model. According to the method, an anchor box does not need to be arranged, the nodule region in the thyroid ultrasound image is efficiently detected, calculation and resource waste related to the anchor box are avoided, the training speed is increased, and the generalization performance of an experimental result is enhanced. The invention further provides a device for detecting the nodules in the thyroid ultrasound image based on deep learning.

The invention relates to a thyroid nodule classification method based on multi-scale feature fusion, and the method is characterized in that the method comprises the steps: 1) obtaining an original thyroid ultrasonic image data set, and processing each ultrasonic image; 2) cleaning the original ultrasonic image data set, and removing images which do not meet requirements to obtain a data set containing 2000 high-quality thyroid nodule ultrasonic images; 3) constructing a thyroid nodule ultrasonic image classification network based on a residual network; 4) replacing the residual module with amulti-scale fusion module; 5), based on a residual network, adding a high-resolution channel; and 6), analyzing a network model classification effect based on multi-scale feature fusion and the high-resolution channel. According to the multi-scale feature fusion and high-resolution channel-based network model classification method, the design is scientific and reasonable, a multi-scale feature andhigh-resolution channel combined mechanism is designed, and the network performance is improved.

The invention discloses a segmentation method for an ultrasonic two-dimensional image of a thyroid nodule. The segmentation method comprises the following steps: carrying out an image enhancement means on the ultrasonic two-dimensional image; the processed two-dimensional image is input into a model based on a U-net structure, the U-net structure comprises n down-sampling operations and n up-sampling operations, the down-sampling operations are composed of convolution modules of a plurality of different convolution kernels, each convolution module is composed of a pooling operation and a convolution operation, and each convolution operation is composed of a pooling operation and a convolution operation; in the convolution operation, convolution kernels with corresponding sizes are adopted to process an input image, and IN operation and Relu operation are added behind the input image; the up-sampling operation consists of a multi-head self-attention mechanism module and a deconvolution module, the deconvolution module consists of convolution with different convolution kernel sizes and bilinear interpolation or transpose convolution, and IN operation and Relu operation are performed after each convolution operation; a segmentation result is enhanced through a loss function and evaluation is carried out; and outputting the segmented ultrasonic two-dimensional image.

The invention relates to an artificial intelligence technology, and discloses a target object position detection method based on a video stream. The method comprises the steps of obtaining the video stream, and performing image framing on the video stream to obtain a framed image set; detecting a target area of the framed image set by using a target area detection model to obtain a target image set; identifying a target object position sequence of the target image set by using a target object position sequence identification model, and deleting target images corresponding to the abnormal target object position sequence from the target image set according to the target object position sequence to obtain a standard target image set; and performing image association on all the target images in the standard target image set, and identifying the position of the target object according to standard target images after image association. In addition, the invention also relates to a blockchaintechnology, and the video stream can be stored in a blockchain. The method can be applied to thyroid nodule position detection. According to the invention, the target object position detection accuracy based on the video stream can be improved.

The invention provides a thyroid nodule TI-RADS grading system and method. Comprising the following steps: denoising a first image data set by applying an image morphological method; performing data enhancement on the first image data set to form a second image data set; training the deep learning network group to form a corresponding first model group; improving the recognition effect of the trained first model group through a domain adaptation method; fusing the first model group through parameter search to obtain a second model. According to the thyroid nodule TI-RADS grading system, the problem of poor automatic classification effect of five factors of thyroid nodules in the prior art is solved.

The invention provides a thyroid image processing method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring a to-be-detected image, wherein the to-be-detected image comprises an image obtained by detecting a thyroid part; inputting the to-be-detected image into a preset deep learning model to obtain a first detection result of the preset deep learning model on the to-be-detected image; and when the first detection result shows that the thyroid nodules exist in the to-be-detected image, inputting a lesion image region representing that the thyroid nodules exist in the to-be-detected image in the first detection result into the tested machine learning model to obtain a second detection result of the machine learning model on the lesion image region. In the scheme, the to-be-detected image is detected in combination with the preset deep learning model and the machine learning model, and the machine learning model does not need to detect the image without the thyroid nodule and the area which is not the lesion image area, so that the operand can be reduced, the operation efficiency can be improved, and the thyroid nodule detection accuracy can be improved.

The invention provides a nodule grading system and method based on thyroid ultrasound images. The method comprises the steps: processing a first database to form a second database suitable for depth model processing; removing unreasonable second thyroid ultrasound images in the second database to form a third database; performing targeted modification on the third thyroid ultrasound image in the third database to form a fourth database; preprocessing a fourth thyroid ultrasound image in the fourth database to form a fifth database; extracting a part of fifth thyroid ultrasound images in the fifth database to form a training database; and through the trained model structure, performing nodule detection on the thyroid ultrasound image in an actual application scene to obtain a nodule grade. The nodule grading system based on the thyroid ultrasound image solves the problem of low accuracy in the thyroid nodule automatic identification and nodule grading process in the prior art.

The embodiment of the invention provides a thyroid nodule edge sign classification method, device and system. The method comprises the following steps: acquiring a thyroid ultrasound image training set; constructing a convolutional neural network model; taking a thyroid ultrasound image in the training set as the input of the convolutional neural network model, taking a classification result as the output of the convolutional neural network model, and training the convolutional neural network model; classifying thyroid nodule edge signs in the thyroid ultrasound image to be processed according to the convolutional neural network model. According to the thyroid nodule edge sign classification method and device, a doctor can be assisted in completing classification of thyroid nodule edge signs, unnecessary puncture operations caused by TI-RADS grading errors due to inaccurate edge sign classification are avoided, and body, money and spirit burdens of a patient are relieved.

The invention relates to the field of digital medical treatment, is applied to the field of intelligent medical treatment, and provides a thyroid nodule aspect ratio calculation method and device, equipment and a storage medium.The method comprises the following steps: acquiring an ultrasonic image of a thyroid nodule; processing the ultrasonic image to obtain a mask of the thyroid nodule shape; extracting the contour of the thyroid nodule according to the mask; calculating the lengths of a main shaft and an auxiliary shaft of the thyroid nodule according to the contour; and calculating the aspect ratio of the thyroid nodule according to the lengths of the main shaft and the auxiliary shaft. By means of the thyroid nodule aspect ratio calculation method, device and equipment and the storage medium, the thyroid nodule aspect ratio can be automatically calculated without depending on judgment of a doctor.

The invention belongs to the technical field of image processing, and particularly relates to a thyroid nodule invasiveness prediction method based on a deep learning segmentation network. The method comprises the following steps: S1, preprocessing a thyroid ultrasound image obtained clinically; S2, constructing a main body structure framework based on a deep learning segmentation network; S3, improving the generative adversarial network model in the main body structure framework; S4, performing accurate semantic segmentation on the thyroid nodule, and counting information of nodule area, aspect ratio and contour rule degree; S5, obtaining a new image data set which only contains nodules after cutting; S6, improving the nonlinear expression ability of the classification network model; and S7, classifying prediction results by using the improved classification network model, and training and updating the classification network model. According to the method provided by the invention, end-to-end automatic auxiliary diagnosis can be realized, and the defects of insufficient accuracy and relatively low detection rate of a traditional detection method are overcome.

The present invention provides an automatic diagnosis system for nodules in thyroid ultrasound images based on multi-scale convolutional neural networks, including: a nodule classification module for thyroid nodules from coarse to fine, an automatic detection module for thyroid nodule regions, and a fine classification module for thyroid nodules ; The convolutional neural network of multi-scale feature fusion extracts the features of the size of different perception areas, so as to combine the local and global information to extract the contextual semantic features of nodules to automatically locate thyroid nodules. The present invention can accurately predict the location of lesions and the probability of occurrence of benign and malignant lesions through the feature extraction of multi-scale coarse-to-fine neural networks and the design of multi-scale fine classification AlexNet with pyramid structure, and can assist doctors in the diagnosis of thyroid lesions. Improving the objectivity of diagnosis, it has the characteristics of good real-time performance and high accuracy rate.

The invention discloses a thyroid nodule segmentation method and device, a storage medium and terminal equipment, and the method comprises the steps: inputting a to-be-segmented thyroid image into a trained first segmentation network model, and determining a nodule positioning region through the first segmentation network model; determining a target image based on the nodule positioning area and the to-be-segmented thyroid image; and inputting the target image into a trained second segmentation network model, and determining a thyroid nodule region through the second segmentation network model. According to the invention, the thyroid nodule is preliminarily positioned based on the first segmentation network model, and then based on the second segmentation network model, segmenting the region of interest corresponding to the thyroid nodule determined based on the preliminary positioning to obtain the thyroid nodule region, so that on one hand, the thyroid nodule region can be automatically obtained, and on the other hand, the accuracy of the thyroid nodule region can be improved through two-stage segmentation.

The invention provides a method for generating a thyroid nodule classification model, and the method comprises the steps: S1, extracting a plurality of thyroid nodule images from a sample database for each nodule type, and adding the images into a corresponding sample subset; S2, constructing a thyroid nodule classification network according to the nodule types, wherein the thyroid nodule classification network comprises a plurality of feature networks, and the feature networks are in one-to-one correspondence with the nodule types; S3, training the feature networks with the same nodule type according to the sample subsets to obtain network parameters corresponding to the feature networks, and configuring the network parameters in a thyroid nodule classification network; and S4, training the thyroid nodule classification network according to the thyroid nodule images in all the sample subsets, keeping network parameters unchanged in the training process, and obtaining a thyroid nodule classification model after the training is completed. The method has the advantage that the problems that an existing thyroid nodule model is too low in recognition accuracy and too high in false positive rate are solved.

The invention discloses a thyroid CT image nodule automatic diagnosis system based on a neural network, and the system sequentially comprises: an image preprocessing module, which carries out the preprocessing of an original thyroid CT image, and carries out the marking of nodule information of the preprocessed image; a image data enhancement module, which is used for expanding the thyroid CT image data set; a nodule semantic segmentation module, which is used for carrying out image semantic segmentation through a neural network to segment nodule parts; an image algorithm optimization module, which enables the output of the semantic segmentation network to be in smooth transition and adapted to the classification network; and a classification prediction module, which is used for carrying out benign and malignant classification judgment on each segmented thyroid nodule by using a hybrid network model. According to the system, end-to-end thyroid nodule diagnosis can be realized, additional image processing and data labeling work on CT images are not needed, and high-accuracy and high-efficiency thyroid nodule automatic identification and benign and malignant classification can be realized.

The invention discloses a thyroid nodule ultrasonic image processing method based on cross-layer sparse atrous convolution, by establishing a novel cross-layer atrous convolution network structure, sparse constraint network, separating sparse atrous convolution layers, and self-adaptive weight adjustment And the loss function of sparse constraints overcomes the poor semantic resolution ability of the existing methods for the nodule area of ​​the thyroid nodule ultrasound image, and the semantic feature extraction of the nodule area is susceptible to similar background interference, and solves the problem of deep learning network in the forward propagation step. Due to the limited receptive field expansion ability, the semantic probability heat map extraction effect of thyroid nodule ultrasound images is not good and other problems.

The invention relates to a thyroid nodule calcification recognition device based on deep learning. The thyroid nodule calcification recognition device comprises an ultrasonic image acquisition module used for acquiring an ultrasonic image data set; a labeling module, which is used for labeling the thyroid nodule boundary, the calcification region and the calcification type of each ultrasonic image; a nodule-of-interest image extraction module, which is used for intercepting a thyroid nodule boundary in each ultrasonic image to obtain a nodule-of-interest image data set; a convolutional neural network construction module, which is used for constructing a convolutional neural network XDNet-11222; a convolutional neural network training module, which is used for training a convolutional neural network XDNet-11222 through the nodule-of-interest image data set; and a thyroid nodule calcification detection module, which is used for carrying out calcification type detection on the input image through the trained convolutional neural network XDNet-11222. According to the invention, the calcification type of the input thyroid nodule ultrasonic image can be effectively identified.

The invention discloses a segmentation method for a thyroid nodule ultrasonic image based on a semantic segmentation network PSPNet, and the method comprises the steps: carrying out the collection andpreprocessing of data, carrying out the manual marking of the thyroid nodule ultrasonic image through combining with a pathological diagnosis result, and dividing each pixel value on the image into three types: thyroid nodule, thyroid parenchyma and other contents; wherein the three types of corresponding pixel values are respectively 3, 2 and 1; training a semantic segmentation network PSPNet; testing a result segmented by the semantic segmentation network PSPNet, and calculating segmentation evaluation indexes such as a cross-parallel ratio and pixel precision; if the test result does not reach the expected standard, adjusting parameters such as the sample number, the loss function, the learning rate and the optimizer of single training of the network, and then training and testing thenetwork until the network reaches the expected standard. In the aspect of segmentation result visualization, smooth parenchyma and nodule edges can be rapidly and specifically segmented, and the segmentation result can be used for further diagnosis.

The invention relates to a comprehensive thyroid nodule prediction method and system and a medium, and belongs to the technical field of information prediction, and the method comprises the steps: obtaining an analog signal transmitted by an ultrasonic probe, and converting the analog signal into digital image information; performing primary data processing on the digital image information to obtain single-frame prediction vector information; performing secondary data processing based on the single-frame prediction vector to obtain prediction result information, and judging whether a thyroid nodule exists in the prediction result information; if the prediction result information exists, establishing a risk assessment model, and importing the prediction result information into the risk assessment model for risk assessment to obtain a risk assessment result report; and performing tracking processing on the risk assessment result report, judging whether the risk assessment result exceeds a preset risk assessment result, and if yes, sending out early warning information.

The invention relates to a thyroid gland detection device. The thyroid gland detection device comprises an image acquisition module which is used for acquiring an ultrasonic image with thyroid nodules; a nodule-of-interest image extraction module which is used for intercepting the thyroid nodule boundary of the ultrasonic image by selecting a plurality of coordinate points of interest to obtain a nodule-of-interest image; and a thyroid gland detection module which is used for carrying out gland detection on the ultrasonic image by taking the nodule in the nodule-of-interest image as a reference. The thyroid gland can be effectively detected.

The invention relates to a thyroid examination and puncture skill training model and relates to the technical field of medical education equipment. The model comprises a high-simulation human body upper body model, a thyroid and thyroid nodule, an electric swallowing action simulation device and a microcomputer monitoring controller. The model is characterized in that a left side thyroid gland and a right side thyroid gland of the high-simulation human body upper body model are respectively provided with circular pits, thyroid benign nodules and thyroid malignant nodules are respectively arranged in the circular pits, inflammatory nodules are arranged at the isthmus of the thyroid gland, microswitches are arranged below the inflammatory nodules, palpation examination can be carried out on the three thyroid nodules, corresponding indicator lamps are turned on, the thyroid inflammatory nodules can be pressed to make a cry; A switch of the microcomputer controller swallowing simulation device is pressed down, swallowing actions can be electrically simulated, thyroid cartilage, thyroid gland and thyroid gland nodules can be driven to move up and down along with the thyroid cartilage, the thyroid gland and the thyroid gland nodules, puncture training and examination can be conducted on the three thyroid gland nodules, and due to the fact that the simulation effect is highly simulated, the teaching quality can be remarkably improved.

The invention discloses an assessment method and device for image information. The method comprises the following steps: extracting feature information from original image information about a thyroid nodule state of a patient; performing feature recognition according to the feature information to determine corresponding text information; performing semantic recognition based on the text information, and determining corresponding structured information; determining a thyroid nodule assessment index of the patient through the structured information by using a preset assessment model. By converting the original image information about the thyroid nodule state of the patient into the text information and structuring the text information, the obtained structured information can be input into the preset assessment model for automatic assessment, so that the thyroid nodule assessment index of the patient is determined, automatic assessment of the thyroid nodule is realized, and and a manual assessment process is avoided.

The invention discloses a thyroid nodule ultrasound image segmentation method, device and system. The segmentation device comprises a data acquisition unit, a model training unit and a segmentation extraction unit. The segmentation system comprises an image segmentation module and a data storage module. A preset first visual segmentation model obtained by introducing a preset hierarchical cascade multi-head self-attention module on the basis of a preset semantic segmentation network Unet is trained and evaluated to obtain a second visual segmentation model, and then a to-be-segmented ultrasonic image is segmented through the second visual segmentation model. According to the thyroid nodule ultrasound image segmentation method, the thyroid nodule ultrasound image segmentation device and the thyroid nodule ultrasound image segmentation system, the segmentation accuracy of the thyroid nodule image is improved.

The invention relates to a thyroid nodule echo analysis device. The thyroid nodule echo analysis device comprises an image acquisition module which is used for acquiring an ultrasonic image with a thyroid nodule; a nodule-of-interest image extraction module which is used for obtaining a nodule-of-interest image according to the ultrasonic image; a thyroid nodule gland detection module which is used for obtaining gland reference according to the ultrasonic image; a detection area generation module which is used for generating a first detection area and a second detection area based on the ultrasonic image; a thyroid muscle and fascia detection module which is used for obtaining muscle and fascia reference according to the first detection area; a thyroid trachea and blood vessel detection module which is used for obtaining a trachea and blood vessel reference according to the second detection area; and a thyroid nodule echo detection module which is used for detecting nodule echo according to the gland reference, the muscle and fascia reference and the trachea and blood vessel reference. The thyroid nodule echo detection method can effectively detect the thyroid nodule echo.

The invention provides ultrasonic diagnosis equipment and a thyroid nodule rating display method thereof. The method comprises the following steps: performing ultrasonic scanning on a thyroid nodule to obtain an echo signal; performing signal processing on the echo signal to obtain ultrasonic data of the thyroid nodule; obtaining a score of a preset score item of the thyroid nodule according to the ultrasonic data; obtaining a severity grade according to the score of each preset score item; a preset score item is used as a classification axis of the radar map, and the radar map is generated and displayed; and displaying the position points of the scores of the preset score items on the radar map, and performing visual expression on the severity rating by adopting corresponding colors in an area formed by connecting the position points. Visibly, the user can see the score of each preset scoring item through the radar map, and can know the severity rating through the color, so that the severity of the thyroid nodule of the patient is clear at a glance.

The invention relates to an intelligent system for automatic segmentation and grading of thyroid nodules, which is characterized in that it comprises: a thyroid ultrasound image database; a thyroid ultrasound image preprocessing module; a thyroid nodule feature extraction module: based on U-Net with ResNet34 as the backbone The segmentation model extracts features from the preprocessed thyroid ultrasound images. The U-Net segmentation model with ResNet34 as the backbone includes a downsampling module, a feature fusion module and an upsampling module; a thyroid nodule segmentation module: used to perform feature extraction The extracted image is semantically segmented to form a thyroid nodule segmentation result map.