19 results about "Liver ct" patented technology

The invention provides a liver tumor early diagnosis method and system based on artificial intelligence, and relates to the technical field of biomedical engineering, and the method comprises the steps: obtaining original information metadata related to the liver, the original information metadata comprises image data original information metadata and biomarker original information metadata, the image data comprises liver CT (computed tomography) and MRI (magnetic resonance imaging) image data, the biomarkers comprise serum tumor marker levels, and the physiological and environmental factors of the patients comprise age, gender, dietary habits and genetic backgrounds of the patients. According to the method, the problems that a tumor in a small and complex area is easily missed or misdiagnosed in the prior art are solved, the path density is predicted through denoising, segmentation and feature extraction of image data preprocessing in combination with a path simulation model based on a graph theory and a Monte Carlo simulation method, and a potential tumor area can be accurately judged.

The invention discloses a neural network model for liver CT image segmentation and a liver cancer classification method based on the neural network model. The neural network model adopts a feature encoder for performing down-sampling feature extraction on the liver CT image; a UNet + + model of a multi-scale feature fusion module connected between the feature encoders and the feature decoders and used for aggregating feature information of different levels, wherein the feature decoders are used for performing up-sampling on the features extracted by the feature encoders to recover spatial dimensions; the feature encoder is integrated with a high-efficiency channel attention module used for enhancing attention to a liver focus key area and inhibiting irrelevant background tissue information and a residual module used for maintaining a network gradient flow; through aggregation and fusion of different levels of feature information, more detail features are reserved when the image space dimension is restored, so that the segmentation precision of the liver and the lesion boundary thereof is remarkably improved, and the segmentation result is more accurate and robust.

The application discloses a liver US-CT medical image conversion method based on a CycleGAN network. The method comprises the following steps: inputting a liver US image x as a training sample into a generator G; the generator G generates a liver pseudo-CT image y' according to the liver US image x; inputting the liver pseudo-CT image y' as a training sample into a generator F; the generator F generates a loopback liver US image x'' according to the liver pseudo-CT image y'; training the generator G; training a discriminator Dy; inputting a liver CT image y as a training sample into the generator F; the generator G generates a liver pseudo-US image x' according to the liver CT image y; inputting the generated liver pseudo-US image x' as a training sample into the generator G; the generator G generates a loopback liver CT image y'' according to the liver pseudo-US image x'; training the generator F; and the discriminator Dx.

The application provides an unsupervised liver tumor CT image segmentation method based on manual features, relates to the technical field of deep learning, and first collects normal liver CT images as a data set, and selects a tumor position in combination with clinical knowledge. A texture similar to real imaging is generated through three-dimensional simple noise binary mask, and histological features are used for morphological modeling. Then, the tumor texture is superimposed with the liver CT image at the selected position to synthesize a new liver tumor CT image. A segmentation model is trained by using the CT image with annotations, and the synthesized image is used for medical image segmentation. Finally, the segmentation results are evaluated by using the Dice coefficient, the Hausdorff distance, the standardized surface distance and the surface distance index. The application is helpful for artificially synthesizing tumor lesion images and generating voxel-level annotations, and provides a new perspective for solving the medical image annotation challenge and promoting liver tumor diagnosis and treatment.

This invention discloses a two-stage controllable liver CT image generation method combining GAN and Diffusion models. The first step involves segmenting a real CT image using a pre-trained nnU-Net to obtain liver and tumor masks. The second step involves training a VQ-VAE model to establish a mapping between pixel space and latent space; training a conditional LDM to perform incisive generation of liver regions: predicting noise under the guidance of multiple hot tags of tumor size, unmasked region features, and the mask. The third step involves rapidly generating a full-frame CT image with global anatomical consistency using GAN; refining liver regions lacking texture detail using the trained LDM; and replacing the liver regions in the full-frame CT image after VQ-VAE decoding. This method combines the global synthesis efficiency of GAN with the local fine-grained generation capability of the diffusion model, effectively improving the realism and diversity of the final synthesized image while significantly reducing computational resources.

The invention relates to a liver tumor segmentation method and device based on adaptive context awareness fusion, and the method comprises the steps: S1, obtaining and preprocessing liver CT data, and dividing the data into a training set, a verification set and a test set according to a proportion; s2, constructing an ACAF-Net (Adaptive Context-Aware Fusion Network), and constructing a double-enhanced dynamic convolution module in the transmission and fusion of the double paths; s3, fusing the double-enhanced dynamic convolution and the double attention weight, and constructing a double-dimensional mixed attention module; s4, fusing the sub-region segmentation error and the dynamic quantization difficulty, and constructing a double-error dynamic balance loss; s5, fusing the double-enhanced dynamic convolution module, the double-dimensional mixed attention module and the double-error dynamic trade-off loss to construct an ACAF-Net model; and S6, completing model training, verification optimization and performance evaluation by using the experimental data set. By using the method, the problems of fuzzy liver tumor boundary, large scale difference and unbalanced segmentation error are effectively solved, the segmentation precision and robustness are remarkably improved, and the method is suitable for automatic segmentation of the CT image liver tumor.

The application belongs to the technical field of CT image processing devices, and particularly relates to a lesion segmentation device based on liver CT images, a preprocessing module pre-processes liver CT images; a segmentation model establishing module establishes a CT image lesion area segmentation model, and then a training module is used to train the CT image lesion area segmentation model; finally, a segmentation module is used to segment the liver and the lesion area of the pre-processed liver CT image by using the trained CT image lesion area segmentation model. The liver CT image is pre-processed to adjust the CT image to a proper size and contrast, 3D convolution is used instead of 2D convolution in the training process of the CT image lesion area segmentation model to retain the connection between slices, [3*3] deep separable convolution is used to extract features, and [1*1] convolution is used to fuse features and adjust the number of channels, thereby effectively improving the accuracy of the segmentation of the lesion area in the CT image and improving the work efficiency of doctors.

The invention provides a liver CT image reconstruction method under a low-dose scanning condition, and relates to the field of image processing, and the method specifically comprises the steps: collecting a standard-dose liver CT image, and generating a low-dose liver CT image; based on a low-dose liver CT image, calculating channel feature difference in a neighborhood range to construct local structure potential energy, generating a structure driving force response value based on potential energy difference, selecting a maximum response direction as a feature aggregation direction, and performing feature fusion in combination with a jump aggregation control factor to obtain a structure aggregation feature map; performing morphological projection density and slope mapping calculation on the structure aggregation feature map, and obtaining a morphological modulation feature map through feature modulation and residual fusion; on the basis of the morphological modulation feature map, combining neighborhood gray difference expansion and multichannel displacement feature analysis, generating a signal modulation factor and performing nonlinear feature regulation and control to obtain an information migration modulation map; and inputting the information migration modulation graph into a reconstruction module for image reconstruction.

The application relates to the technical field of liver segmentation, in particular to a liver image segmentation method based on multi-head attention feature fusion, which comprises the following steps: establishing a liver image segmentation model based on an Unet network structure, collecting liver CT images and corresponding labeled region images; performing edge detection operator processing on the liver CT images to obtain a liver region contour image; performing binaryzation processing on the liver CT images to obtain a liver region binaryzation image; obtaining a liver region rough segmentation image through a segmentation network; after adjusting the liver CT images, the labeled region images, the liver region contour image, the liver region binaryzation image and the liver region rough segmentation image to the same size, the images are combined to obtain a combined image; training the liver image segmentation model by using the combined image set, calculating a loss by using a loss function, optimizing model parameters according to the loss until convergence; inputting a to-be-processed liver CT image into the trained liver image segmentation model to obtain a liver segmentation image; and the application can provide a more accurate segmentation image.

The application discloses a liver CT image continuity discontinuous dynamic imaging information fusion method, which combines a machine learning algorithm, filters and feature segments images by extracting and identifying feature points, obtains a dynamic change curve graph and a standard floating line, constructs a maximum loss difference value and a minimum loss difference value of the dynamic change curve graph, establishes a loss difference value allowable range, obtains minimum values and maximum values in a dynamic change period, judges threshold difference based on the maximum values, realizes liver CT image fusion image probability analysis, and solves the problems that in the prior liver change diagnosis and treatment process, doctors have heavy workloads, evaluation efficiency is relatively low, and analyzed lesion information is not comprehensive enough, so that key feature points are easily ignored, and diagnosis and treatment difficulty is increased.

The application discloses a liver tumor image segmentation method based on a CNN-Transformer hybrid architecture, and comprises the following steps: pre-processing a liver CT image to reduce the influence of image acquisition differences on a model; a multi-stage hybrid encoder effectively combines a convolutional neural network and a Transformer network, extracts local details, takes into account the dependency relationship between elements, and obtains multi-scale features; a feature enhancement decoder performs up-sampling through a convolutional neural network, realizes multi-scale feature fusion through an improved skip connection, and finally outputs a segmentation probability graph with a size of 512*512*3; and a post-processing module converts the segmentation probability graph into a segmentation result that can be directly used for surgery planning. The application constructs a powerful segmentation model that can grasp local details and understand global context, improves calculation efficiency, and reduces the consumption of computing power resources.

The invention discloses a liver CT image segmentation method, system and device, a medium and a product, and relates to the field of image processing, and the method comprises the steps: inputting a liver CT image in a case into an encoder, extracting and gradually downsampling image features, and flattening the highest-layer semantic features of the image features into a sequence; based on a Transform module, generating a preliminary hidden feature; a decoder is utilized to decode an image with preliminary hidden features for the first time, meanwhile, a cross fusion module is combined to carry out cross fusion processing on coding features from different convolutional layers of an encoder, the coding features are fused with up-sampling features, fused features are determined, edge slice features are determined by utilizing a slice sensing adjustment module, and the edge slice features are extracted by utilizing an image extraction module. Determining non-edge slice features by using a reverse attention module; the edge slice features and the non-edge slice features are reversely fused, the fused feature image is decoded for the second time, the segmentation result of the liver CT image is generated, and the performance of continuous CT image segmentation is improved.

A method based on an r-IBS-introduced neural network for automatically determining the liver's condition from three common diseases—hepatocellular carcinoma (HCC), intrahepatic cholangiocarcinoma (ICC), and fatty liver (FAT)—using liver CT images includes the following steps: First, construct an r-IBS-based training classifier. Use the classifier to determine the disease type for each liver CT image. Acquire liver CT images and process them using a snake model. Input the images into the trained classification neural network model. The neural network outputs the classification results for the aforementioned diseases.

The application provides an image automatic segmentation method based on a multi-level multi-attention MLMA-UNet network, solves the problems of high calculation complexity and low segmentation performance in the prior art, and comprises the following steps: step one: acquiring a liver CT image dataset and pre-processing the CT image; step two: constructing a multi-level feature recalibration network segmentation model for the liver and tumors, training the model by using a training set, extracting global and local features by using multi-level extraction, and recalibrating the channel response of the aggregated multi-level features; step three: adjusting the parameters of the multi-level feature recalibration network segmentation model and training the model multiple times, obtaining a stably converged model when the loss function of the model stably converges, testing the trained model by using a test set, constructing a liver and tumor detection network, obtaining liver and tumor segmentation results, and evaluating the network performance by using statistical metrics.

An early liver cancer diagnosis model based on CapsNet involves image processing. Features are extracted from the input enhanced liver CT image and processed through a ResNet module using convolution, regularization, activation functions, and max pooling. The output features are captured in the PrimaryCaps layer to obtain primary image features, reducing dimensionality and refining the features. An additional PrimaryCaps layer is added between the PrimaryCaps and DigitCaps layers to identify and integrate local features. After classification by the DigitCaps layer, a digital capsule layer is included for dynamic routing, and a Squash operation is performed to obtain a feature map composed of 8-dimensional vectors. E-CapsNet is designed by fusing the deep residual learning framework of ResNet and the high-level feature representation capabilities of CapsNet to improve the accuracy and efficiency of early liver cancer diagnosis.

This invention proposes a method for liver CT image reconstruction under low-dose scanning conditions, relating to the field of image processing. The specific steps include: acquiring standard-dose liver CT images and generating low-dose liver CT images; based on the low-dose liver CT images, calculating channel feature differences within a neighborhood to construct local structural potential energy, generating structural driving force response values ​​based on potential energy differences, selecting the direction of maximum response as the feature aggregation direction, and combining it with a jump aggregation control factor for feature fusion to obtain a structural aggregation feature map; calculating morphological projection density and slope mapping on the structural aggregation feature map, and obtaining a morphological modulation feature map through feature modulation and residual fusion; based on the morphological modulation feature map, combining neighborhood grayscale difference expansion and multi-channel displacement feature analysis to generate a signal modulation factor and perform nonlinear feature modulation to obtain an information migration modulation map; and inputting the information migration modulation map into a reconstruction module for image reconstruction.

The invention discloses a liver temperature distribution optimization processing method and device based on personalized magnetic thermal ablation, and the method comprises the steps: inputting to-be-detected liver CT data into a liver multi-tissue segmentation network model, and obtaining segmented tumor and blood vessel data; the model is trained and generated by using a synthesized liver CT data set, the data set is synthesized by inputting liver mask data containing blood vessels and tumors into a condition diffusion model, and the diffusion model is trained and generated by using a relation sample data set between the liver mask data containing blood vessels and tumors and the synthesized liver CT data; reconstructing a personalized liver digital model according to the segmented tumor and blood vessel data; according to the liver digital model and a multi-physical field simulation optimization system, optimizing temperature distribution of different magnetic hyperthermia treatment schemes in the liver; the distribution is used to generate an optimal combination of magnetic hyperthermia parameters. According to the method, different tissues in the reconstructed liver model can be accurately segmented, so that the liver temperature distribution is accurately evaluated.

The invention belongs to the field of image processing, particularly relates to a semi-supervised liver CT image segmentation method and system based on cooperative training, and aims to solve the problem of inaccurate medical image segmentation. According to the method, a teacher-student double-model cooperative training framework is adopted, and a large number of unlabeled CT images which are easy to obtain can be effectively utilized for learning. A consistency loss function is constructed based on a consensus region and a divergence region, so that different regions can learn mutually. And then a supervision loss function and a consistency loss function of the coordinated training framework are mutually combined to construct a total loss function to train a multi-branch prediction network to carry out liver tissue area image segmentation, and the accuracy and segmentation precision of a segmentation model are further improved. According to the method, the accuracy and the precision of CT image segmentation are improved, and particularly, the segmentation precision of fully supervised learning can be achieved or even exceeded under the condition that only a small amount of annotated data exists.

The application discloses a CT image auxiliary detection method and system for liver focal lesions, and the method comprises the following steps: receiving and screening multi-dimensional liver CT image data, generating a standardized DICOM data set based on preset layer thickness threshold, resolution threshold and artifact filtering rules; performing two-stage lesion positioning analysis on the standardized DICOM data set, and outputting a lesion positioning atlas containing lesion coordinates, size and average CT value; starting an interactive image optimization engine based on the lesion positioning atlas, responding to the window width and window level adjustment instructions and measurement tool operations of a doctor in real time, generating a dynamically enhanced lesion visual field and synchronously updating lesion parameters; integrating the visual field data and a clinical rule base to generate a structured report, automatically filling the image findings field by checking the lesion list, and outputting a diagnosis proposal conforming to the DICOM standard through a graphic-text report module. By using the embodiment of the application, the standardization and intelligentization of auxiliary diagnosis can be realized, and the precision and efficiency of liver lesion detection are improved.