125 results about "Nodule detection" patented technology

The invention relates to an automatic detection system for a pulmonary nodule in a chest CT (Computed Tomography) image. The system provides improvements for the problems of large calculated amount of computer aided software, inaccurate prediction and few prediction varieties. The improvement of the invention comprises the steps of acquiring a CT image; segmenting a pulmonary tissue; detecting a suspected nodular lesion area in the pulmonary tissue; classifying nidi based on a nidus classification model of deep learning; and outputting an image mark and a diagnosis report. The system has high nodule detection rate and relatively low false positive rate, acquires an accurate locating, quantitative and qualitative result of a nodular lesion and a prediction probability thereof. The end-to-end (from a CT machine end to a doctor end) nodular lesion screening is truly realized, the accuracy and operability demands of doctors are satisfied and the system has wide mar5ket application prospects.

The invention relates to a lung nodule detection method based on a three-dimensional convolution neural network, which adopts a network structure of a feature pyramid and an attention mechanism,. Themethod integrates features of low-level and high-resolution and high-level abstract features, and enables the network to concentrate attention on a target area. The method has the effects of using three-dimensional convolution neural network detection method, end-to-end detection of pulmonary nodules, and reducing the time overhead. Compared with the traditional methods, the method can improve therecall rate and average accuracy of nodule detection.

The invention relates to a lung nodule detection method and device, a computer device and a storage medium. The method comprises the following steps segmenting the lung CT image through a three-dimensional convolution neural network segmentation model, and segmenting a lung region image; detecting suspicious nodules from the lung region image through a three-dimensional U-Net detection model; classifying the suspicious nodules through a three-dimensional dichotomous network, and removing the fake nodules. According to the lung nodule detection method and device, the computer device and the storage medium, the lung CT image is segmented through the three-dimensional convolution neural network segmentation model; the lung region image is segmented, the segmentation speed is high, and the speed of subsequent detection of the lung nodules is accelerated; meanwhile, the invention can be applied to the segmentation of the lung region of all lung CT images.

The invention discloses a CT lung nodule detection system based on 3D-Unet. The system includes a CT image input and preprocessing module, a candidate nodule detection module and a false-reporting elimination module which are sequentially connected. The CT image input and preprocessing module is used for reading a chest CT graph, obtaining spacing and origin information of the CT graph, and carrying out lung volume segmentation on the CT graph. The candidate nodule detection module is used for inputting the image, of which preprocesing is completed, into a Unet network, and obtaining a location of a candidate lung nodule. The Unet network includes sixteen residual blocks, two path units, four maximum-pooling units, one preparation unit, one probabilistic neuron failure unit and one outputunit. According to the system, automatic detection of the lung nodule is realized, problems of high doctor workloads and high misdiagnosis rates in lung nodule diagnosis are solved, an effect of detecting the lung nodule by the 3D-Unet is realized, more contextual semantic information in the CT image is utilized, and detection accuracy is higher.

A method of generating a pulmonary nodule image from a chest radiograph. The method includes the steps of: producing a map of a clear lung field; removing low frequency variation from the clear lung field to generate a level image; and performing at least one grayscale morphological operation on the level image to generate a nodule-bone image. Pulmonary nodules can be detected using the nodule-bone image by the further steps of: pulmonary nodules from a chest radiograph. The method includes the steps of: identifying candidate nodule locations in the nodule-bone image; segmenting a region around each candidate nodule location in the nodule-bone image; and using the features of the segmented region to determine if a candidate is a nodule.

An incremental greedy method to feature selection is described. This method results in a final classifier that performs optimally and depends on only a few features. Generally, a small number of features is desired because it is often the case that the complexity of a classification method depends on the number of features. It is very well known that a large number of features may lead to overfitting on the training set, which then leads to a poor generalization performance in new and unseen data. The incremental greedy method is based on feature selection of a limited subset of features from the feature space. By providing low feature dependency, the incremental greedy method 100 requires fewer computations as compared to a feature extraction approach, such as principal component analysis.

A method for classifying pulmonary structures in digitized images includes providing approximate target structure locations of one or more target structures in a digitized 3-dimensional (3D) image, fitting an anisotropic Gaussian model about said approximate target locations to generate more precise 3D target models and center locations of said one or more target structures, warping each said 3D target model into a 3D sphere, constructing a bounding manifold about each said warped 3D sphere, and identifying clusters on said bounding manifold wherein said one or more target structures are classified.

The invention discloses a lung nodule detection method based on a 2D convolutional neural network. The method improves the detection accuracy by detecting suspected pulmonary nodules and reducing false positives. Meanwhile, the whole detection process can be automatically completed, and the detection efficiency of the lung nodules is accelerated.

The invention discloses a CT image pulmonary nodule detection method based on an improved Faster R-CNN framework. The method comprises the following steps: 1, collecting the CT images of the chest ofthe patients with pulmonary nodule symptoms, and marking the position of the pulmonary nodule as a training sample set; 2, constructing a candidate nodule detection network, training that candidate nodule detection network with a training sample set, determining network parameters, and obtaining a candidate nodule detection model; 3, constructing a false positive rejection classification network of candidate nodules, and training the false positive rejection classification network of candidate nodules with a training sample set to obtain a false positive rejection classification network modelof candidate nodules; 4, inputting a CT image to be detected into a candidate nodule detection model to obtain a position of the candidate nodule; inputting the position of the candidate nodules intoa false positive rejection classification network model of the candidate nodules, eliminating false positive rejection, and obtaining the detection result of the lung nodules. This method is more suitable for the detection of small pulmonary nodules.

The invention discloses a medical image lung nodule detection method based on an anti-network, which mainly solves the problem of low detection precision due to shading on medical image lung nodules in the prior art. The method comprises the steps of: 1) preprocessing a medical image to acquire a sample data set; 2) scaling and cutting the sample data set, and adding Gaussian noise to all processed samples to form an expanded sample data set; 3) superimposing a Faster-RCNN detector with an anti-space discarding network (ASDN) to construct a novel detector based on the anti-network; 4) trainingthe novel detector with the sample data set to obtain a trained novel detection model; and 5) performing lung nodule detection on a test data set with the trained detection model to obtain a lung nodule detection result in each medical image. The method improves the detection precision of medical image lung nodules and can be used in a computer aided diagnosis system.

The invention discloses a lung cancer screening method based on a neural network, and the method comprises the steps: data preparation: mainly completing the importing of data from a hospital data system and the calibration of data at the stage; preprocessing of the data, namely preprocessing the data and segmenting a lung region in the stage; constructing and training of a three-dimensional RPN network, namely training the RPN network by using the acquired data and the calibration result; training networks suitable for different resolutions, and constructing a combination strategy; carrying out pulmonary nodule detection on the input three-dimensional CT image by using the trained model, and outputting predicted pulmonary nodule information including a central point three-dimensional coordinate and the diameter of the pulmonary nodule; constructing a neural network based on an attention mechanism according to the three-dimensional coordinates and the diameter of the central point given in the step 5, and judging the property of the pulmonary nodule. According to the method, candidate nodule detection is directly carried out on three-dimensional input, so that candidate detection is more comprehensive and reliable.

Embodiments of the invention provide a nodule detection method and device, and relate to the technical field of machine learning. The method comprises the following steps of: obtaining a nodule image,extracting feature images, with different sizes, of the nodule image by adoption of a nodule detection model, and determining nodule recognition boxes from the feature images, wherein the nodule detection model is determined after training a plurality of nodule images of marked nodules by adoption of a 3D convolutional neural network; and determining nodules of the nodule images according to thenodule recognition boxes determined from the feature images. According to the method, the feature images, with different sizes, of the nodule images are extracted, and nodules in each feature image are recognized, so that the nodules with big sizes can be detected and the nodules with small nodes also can be detected, thereby improving the nodule detection precision. Compared with the method of manually judging whether nodules exist in nodule images or not, the method for automatically detecting nodules by adoption of the nodule detection model is capable of effectively improving the nodule detection efficiency.

This application relates to a tissue nodule detection and tissue nodule detection model training method, apparatus, device, storage medium and system. The method for training a tissue nodule detection model includes: obtaining source domain data and target domain data, the source domain data comprising a source domain image and an image annotation, the target domain data comprising a target image, and the image annotation being used for indicating location information of a tissue nodule in the source domain image; performing feature extraction on the source domain image using a neural network model to obtain a source domain sampling feature, performing feature extraction on the target image using the neural network model to obtain a target sampling feature, and determining a model result according to the source domain sampling feature using the neural network model; determining a distance parameter between the source domain data and the target domain data according to the source domain sampling feature and the target sampling feature, the distance parameter being a parameter describing a magnitude of a data difference between the source domain data and the target domain data; determining, according to the model result and the image annotation, a loss function value corresponding to the source domain image; and training the neural network model to obtain a tissue nodule detection model by iteratively reducing a combination of the loss function value and the distance parameter. In this way, the detection accuracy can be improved.

The invention discloses a CT image lung nodule detection method, apparatus and device and a readable storage medium. The CT image lung nodule detection method comprises the following steps: when receiving a medical record processing instruction, obtaining a target CT image corresponding to the medical record processing instruction; acquiring and adjusting display parameters of the target CT imageto obtain an adjusted image; through the pre-stored three-dimensional convolution neural pixel segmentation network and the three-dimensional convolution neural network classifier, analyzing the adjustment image to determine the nodular region, obtaining and outputting the determined nodular region of the adjustment image, and obtaining the first analysis information of the determined nodular region. The invention solves the technical problems of low detection efficiency and low accuracy of the existing artificial lung nodules.

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.

A method of detecting a nodule in a three-dimensional scan image comprises calculating a three-dimensional sphericity index for each point in the scan image (310-330), applying a high sphericity threshold to the sphericity index (340) to obtain a candidate nodule region, and then performing region-growing (350) from the candidate region using a relaxed sphericity threshold to determine an extended region including less spherical parts connected to the candidate region. Optionally, spherical filtering may be applied to the image by matching the spherical filter to the extended region.

The invention discloses a time series based treatment effect evaluation method. The time series based treatment effect evaluation method comprises the steps of: acquiring lung CT image data, wherein the lung CT image data is composed of a plurality of lung sections; performing pre-processing on lung CT image data; designing a lung nodule detection model according to the pre-processed lung CT imagedata; designing a lung nodule classification model; designing a time axis based lung cancer prediction model according to the lung nodule detection model; and training the lung nodule detection modelaccording to the number of samples per batch. Based on the deep learning, a time series is added, the method dynamically analyzes the changes of the lung nodules, so that a doctor can more accuratelydiagnose a treatment effect of the lung cancer and predict the early recurrence. The diagnosis time of the algorithm is far less than the doctor's manual diagnosis time, and the important steps of the doctor's diagnosis can be assisted in a diagnosis process, which can save a lot of time and energy of the doctor.

Image processing techniques may include a methodology for normalizing medical image and/or voxel data captured under different acquisition protocols and a methodology for suppressing selected anatomical structures from medical image and/or voxel data, which may result in improved detection and/or improved rendering of other anatomical structures. The technology presented here may be used, e.g., for improved nodule detection within computed tomography (CT) scans. While presented here in the context of nodules within the lungs, these techniques may be applicable in other contexts with little modification, for example, the detection of masses and/or microcalcifications in full field mammography or breast tomosynthesis based on the suppression of glandular structures, parenchymal and vascular structures in the breast.

The invention discloses a lung nodule detection model construction method, and aims to solve the technical problem that the existing detection method is low in efficiency, inconvenient to operate andincapable of being applied to large-scale samples. The method comprises the following steps: (1) screening out a lung nodule training sample and carrying out the data enhancement; (2) constructing a Caffe depth learning framework, wherein the configuration of the model is completed by utilizing a Faster R-CNN network, and a VGG-16 network structure is introduced; (3) performing the applicable lungnodule image improvement on the detection model; (4) adjusting the network parameters to enable the model to be converged; (5) optimizing by utilizing a loss function; (6) training the designed network, so as to obtain an image recognition network with the function of detecting lung nodules. The method has the advantages of being high in detection speed and high in precision, being applied to large-scale samples, being capable of self-updating iteration, and being high in intelligent degree.

The invention discloses a prediction method for an early-stage lung cancer based on deep learning. The method comprises the following steps of acquiring lung CT (Computed Tomography) image data, wherein the lung CT image data consist of a plurality of lung slices; performing a pretreatment operation on the lung CT image data; designing a lung nodule detection model; designing a lung nodule classification model; and training the lung nodule detection model according to a sample number in each batch. According to the prediction method for the early-stage lung cancer based on deep learning, the original CT data are converted into model trainable data through a preprocessor, the trained model is deployed in a server, a lung nodule of a patient whose CT data are uploaded is detected by use of the model of the server, a nodule position and a nodule size are output, the benignness and malignancy of the nodule are detected and prediction is performed on the nodule state of next check.