538results about How to "Improve segmentation efficiency" patented technology

The invention provides an automatic segmentation method for an MRI (Magnetic Resonance Imaging) image brain tumor based on a full convolutional network. The method comprises multi-mode MRI image preprocessing of the brain tumor, construction of a full convolutional network model, network training and parameter optimization as well as automatic segmentation of a brain tumor image, specifically, the segmentation of the MRI image brain tumor is converted into a pixel-level semantic annotation problem and differential information emphasizing different modes of MRI, two-dimensional whole slices of four modes FLAIR, T1, T1c and T2 are synthesized into a four-channel input image, the convolutional layer and the pooling layer of the trained convolutional neural network are base feature layers, three convolutional layers equal to a full connection layer are added behind the base feature layers to form a middle layer, the middle layer outputs rough segmentation images corresponding to semantic segmentation types in quantity, and a de-convolutional network is added behind the middle layer and used for interpolating the rough segmentation images to obtain a fine segmentation image having the same size as the original image. The method does not need manual intervention, effectively improves the segmentation precision and efficiency, and shortens the training time.

The invention discloses a computerized auxiliary diagnosing method for malignant tumor in early stage based on deep learning algorithm. According to the invention, a deep convolution neural network is used to extract the high-level characteristics of an image layer by layer so that the obtained characteristic dimensions become smaller and smaller. To realize the dimension matching between an outputted probability predicting image and a pre-segmented image, the method uses the reverse convolution network to expand the dimensions of the characteristic image so as to obtain a dimension-consistent probability predicting image. The network generated probability predicting image extracts the binarization of the predicting image through the training of a softmax classifier to finally obtain the segmentation result of the prostate tissue. To some extent, the method effectively increases the accuracy and efficiency for doctors to diagnose prostate diseases.

The invention provides a complex background real-time alternating method based on background modeling and energy minimization, comprising the steps of: firstly, segmenting a current frame image into a foreground pixel set, a background pixel set and an unknown label pixel set by using a provided fusion background model; secondly, constructing a target energy function according to the color and contrast information under a dynamic diagram cutting frame, designing a data item of the fusion background model based on time continuity information and a contrast smooth item based on a local two-value mode, solving two-value labels of all pixels by using a diagram cutting algorithm minimization energy function; and seamlessly fusing segmented foreground targets into a virtual background by adopting postprocessing methods of boundary smooth, alpha value estimation and the like. Experimental results show that the method can better segment the targets in the complex background in real time and really synthesize the segmented targets into the virtual background. The invention has the characteristics of automatically segmenting the complex background in real time and replacing the segmented complex background into the virtual background to obtain a virtual effect sequence diagram with high quality.

The invention discloses an interactive type image segmentation and fusion method based on a grabucut algorithm. The interactive type image segmentation and fusion method based on a grabucut algorithmcomprises steps of adopting a pyramid down-sampling to reduce a resolution of an image, using few typical pixel points to estimate a GMM parameter, using an interactive type technology to mark a foreground and a background of the image, performing watershed segmentation, performing modification on a mask image obtained from segmentation and transmitting the modified the mask image to a grabucut for fine segmentation, performing morphological processing on the segmented image, continuously performing pixel transformation and morphological processing on a processed image to obtain a tripartite graph, using the trimap graph and a source graph as inputs, using a shared matting algorithm to obtain an alpha channel image of the foreground image, using the alpha channel image, the source image and a new background image as inputs and using an alpha blending transparent mixed algorithm to obtained a fused image.

The invention discloses an image type fire flame identification method. The method comprises the following steps of 1, image capturing; 2, image processing. The image processing comprises the steps of 201, image preprocessing; 202, fire identifying. The fire identifying comprises the steps that indentifying is conducted by the adoption of a prebuilt binary classification model, the binary classification model is a support vector machine model for classifying the flame situation and the non-flame situation, wherein the building process of the binary classification model comprises the steps of I, image information capturing;II, feature extracting; III, training sample acquiring; IV, binary classification model building; IV-1, kernel function selecting; IV-2, classification function determining, optimizing parameter C and parameter D by the adoption of the conjugate gradient method, converting the optimized parameter C and parameter D into gamma and sigma 2; V, binary classification model training. By means of the image type fire flame identification method, steps are simple, operation is simple and convenient, reliability is high, using effect is good, and the problems that reliability is lower, false or missing alarm rate is higher, using effect is poor and the like in an existing video fire detecting system under a complex environment are solved effectively.

The invention provides a cloud atlas segmentation method based on a FCN and a CNN, and belongs to the field of image segmentation of computer vision. The method is characterized in that firstly, a near neighborhood of each pixel point in a cloud atlas is correspondingly clustered through an ultra-pixel, the cloud atlas is input to fully convolutional networks FCN32s and FCN8s with different step lengths, and pre-segmentation results of the cloud atlas are realized; in a FCN32s result map, a black area is a part of a non-cloud area in the cloud atlas, and in a FCN8s result map, a white area is a part of a cloud area in the cloud atlas; and remaining uncertain areas are grey areas which need to be determined through a deep convolutional neural network (CNN), key pixels in an ultra-pixel area need to be selected to represent characterizes of the ultra-pixel area, and whether the characteristics of the pixels are cloud or non-cloud is determined through the CNN network. According to the method, the precision is equivalent to that of MR-CNN and SP-CNN, and the speed is increased by 880 times compared with MR-CNN and increased by 1.657 times compared with SP-CNN.

The embodiment of the invention provides an image foreground and background segmentation network model training method, an image foreground and background segmentation method, an image processing method and device, and a terminal device. The image foreground and background segmentation network model training method comprises the following steps: obtaining the eigenvectors of a sample image to be trained; performing convolution on the eigenvectors to obtain the convolution results of the eigenvectors; magnifying the convolution results of the eigenvectors; determining whether the magnified convolution results of the eigenvectors satisfy the convergence conditions or not; if so, completing the training of the convolutional neural network model used for segmenting the foreground and background of the image; and if not, adjusting the parameters of the convolutional neural network model according to the convolution results of the amplified eigenvectors and performing iteration training on the convolutional neural network model according to adjusted parameters of the convolutional neural network model until the convolution results satisfy the convergence conditions. By means of the image foreground and background segmentation network model training method, the training efficiency of the convolutional neural network model is improved, and the training time is shortened.

The invention discloses an image enhancement and partition method, which comprises the following steps of I, image enhancement: a processor and an image enhancement method based on fuzzy logic are adopted for performing enhancement processing on an image needing to be processed, and the process is as follows: i, converting from an image domain to a fuzzy domain: mapping the grey value of each pixel point of the image needing to be processed to a fuzzy membership degree of a fuzzy set according to a membership function (shown in the specifications); ii, performing fuzzy enhancement processing by utilizing a fuzzy enhancement operator in the fuzzy domain; iii, converting from the fuzzy domain to the image domain; II, image partition: partitioning a digital image subjected to enhancement processing, i.e. a to-be-partition image, according to an image partition method based on a three-dimensional fuzzy division maximum entropy. The method is simple in steps, reasonable in design, convenient to realize, good in processing effect and high in practical value, and an image enhancement and partition process can be simply, conveniently and quickly completed with high quality.

The invention discloses an image segmentation method and device and belongs to the field of image processing. The image segmentation method includes the following steps: establishing a significance model of an image; according to the significance model, obtaining foreground sample points and background sample points in the image; according to the significance model and the foreground sample points and the background sample points, establishing a foreground and background classification model; and according to a preset image segmentation algorithm, segmenting the image, wherein the preset image segmentation algorithm uses the foreground and background classification model and edge information between pixel points to segment the image. Through automatic determination of the foreground and background sample points in combination with the significance model, the foreground and background classification model is established and the foreground and background classification model is used to realize image segmentation. Therefore, problems, which exist in related technologies, that the foreground sample points and the background sample points must be selected roughly manually and the segmentation efficiency is comparatively low when a large quantity of images are segmented are solved so that effects of realizing automation selection of samples and improving the classification precision and segmentation efficiency are achieved.

The invention discloses a three-dimensional structural description method of hepatic vascular system, which belongs to the field of medical liver division. The method comprises the following steps of: firstly, dividing a hepatic image and a hepatic vascular image, structurally analyzing the hepatic vascular three-dimensional image, skeletonizing the liver and the blood vessel, and marking the skeletonized voxel points; next, storing the marked voxel points to data structure of multi-branch tree; then, distinguishing main and branch blood vessels according to master slave relation of the voxelpoints in the multi-branch tree; finally, dividing the liver body into multiple liver segments according to relation between the blood vessels in the hepatic vascular system. By the three-dimensionalstructural description method of hepatic vascular system, accurate division of liver is realized so that guidance is provided for accurate removal of liver.

The invention discloses a framework for improving the effect of semantic segmentation model based on transfer learning, which comprises the following contents: 1) introducing the transfer learning into the semantic segmentation field, so that the fast semantic segmentation network can improve the segmentation effect of the student model through the teacher model; 2) consistency mapping being proposed to measure the contour and texture information of the teacher-student model, and a consistency loss function being constructed to make the fast semantic segmentation better in detail. 3) using a teacher model and a conditional random field (CRF) model to generate auxiliary label for unlabeled data, and adding data to a training set to improve that generalization ability and segmentation effectof the model. The invention improves the accuracy rate of the fast semantic segmentation model without introducing additional model parameter and reducing the model speed.

The invention discloses a weed image segmentation method under the rape field environment. Multiple weed/rape RGB image samples are randomly acquired in the rape field; a visual attention model is established, the color characteristics, the brightness characteristics and the direction componential characteristics are extracted, each characteristic graph is acquired and each characteristic channelsaliency map is generated so that a total saliency map is acquired and the area of interest is acquired; the shape characteristics and the texture characteristics of the area of interest are extractedto perform support vector machine classification training so as to acquire the rape area; and the miscellaneous image samples and all the rape area images are fused so as to acquire the final inter-strain weed area distribution information. The area of interest is acquired through fusion of the improved visual attention model with combination of the region growth algorithm, and the whole algorithm process does not require grayscale transformation or threshold segmentation so that the processing link and the computing amount can be reduced; and the segmentation efficiency is further enhanced by extracting the characteristic parameters of the area of interest and support vector machine classification model judgment so that weed image segmentation under the background of the rape field can be realized.

The invention provides an automatic retinal blood vessel segmentation method for clinical diagnosis of glaucoma. Through the method, the influence of bright regions such as optic disc and exudate is eliminated by fusing five models depending on different image processing technologies, namely a matched filter, a neural network, multi-scale line detection, scale space analysis and morphology. At thesame time, massive data is not needed to establish a retinal blood vessel segmentation model, so the method greatly reduces the amount and complexity of data to be processed, is easy to realize, andcan effectively improve the efficiency of retinal blood vessel segmentation. The method also utilizes the region growth method and the gradient information to carry out iterative growth on the background and the blood vessel region on the basis of a multimodal fusion result, the segmentation results exhibit better continuity and smoothness, more retinal vascular details and more complete retinal vascular network can be kept, and thus the method effectively assists ophthalmologists in diagnosing diseases and lightens the burden of ophthalmologists.

The invention provides an automatic segmentation method based on self-adaptive external force level set for magnetic resonance images (MRIs) of soft tissue and a realization method thereof. The automatic segmentation method comprises the steps that object region tissue satisfying a gray threshold is selected under the circumstance of interactive operation; gradient information of a smoothened image is calculated; a curve evolution guiding function is set; a boundary stopping function is set; and if an evolutional curve exceeds the boundary of an object region, parameters of the boundary stopping function are adjusted, then curve evolution is re-conducted and the iteration is conducted again until the evolutional curve is converged within the boundary of the object region. The automatic segmentation method has the characteristics of high real-time performance, high operation efficiency, capability of segmenting multiple discrete regions at the same time, capability of accurately recognizing fuzzy boundaries of the soft tissue of a human body, high segmentation precision, clear image detail characteristics, high intelligent degree, no need of manual intervention, stable and reliable operation, and the like.

The invention discloses a Chinese character segmentation method for off-line handwritten Chinese character recognition. The method is characterized by comprising the following steps: (1) a Chinese character image to be recognized is pre-processed (including binarization of the image); (2) the Chinese character image is roughly segmented on the basis on projection analysis, the non-adherent characters are segmented into single characters and the adherent characters are integrally segmented; (3) the average height of the non-adherent character is acquired; (4) the characters segmented in the step (2) are judged according to the average height of the non-adherent characters, which is acquired in the step (3), so as to obtain an adherent character string set; and (5) each adherent character string in the adherent character string set is re-segmented on the basis of minimum weighting segmentation path, so as to achieve segmentation of the adherent characters. The method is effectively self-adaptive to segmentation of both adherent Chinese characters and non-adherent Chinese characters, and has higher segmentation accuracy and efficiency.

The invention discloses a satellite image segmentation method based on residual network and U-Net segmentation network. The satellite image segmentation method comprises the following steps: constructing a residual network ResNet 34; constructing U-Net segmentation network; constructing a training sample set; training the residual network ResNet 34; training the U-Net segmentation network; inputting the satellite image to be segmented into the residual network ResNet 34 for binary classification, and judging that a ship target is included; using the U-Net segmentation network to perform binarysegmentation on the positive samples in the classification result; for the negative samples in the classification result, directly outputting a single-value mask graph. According to the method, the satellite image is subjected to binary classification by using the residual network ResNet 34, and the U-Net segmentation network is used to only segment the positive samples in the classification result, and an SE-ResNet module is embedded in the U-Net segmentation for extracting finer segmentation masks and the satellite image segmentation method is high in real-time performance and segmentationprecision.

The invention relates to a blood vessel ROI dividing method based on an intravascular ultrasonic image. The method includes the steps that a lumen area and a lumen membrane outline of a blood vessel are divided firstly, the initial outline of a parameter active outline model is obtained by positioning the center of the lumen area, then the middle and outer membrane outline curve of the blood vessel is obtained through convergence, and priori knowledge of the lumen area information is used sufficiently for extracting a middle and outer membrane; an inner area is taken as ROI for the middle and outer membrane outline curve, and blood vessel plaques are divided by minimizing a movable outline model globally. Visualization of outline information of the middle and outer membrane and the plaques of the lumen membrane of the blood vessel ROI is achieved, in comparison with an IVUS image dividing method based on statistics, a complex statistical modeling process of the IVUS image dividing method is abandoned, and the dividing result is not affected by IVUS image artifacts and plaque features; the step of dividing the initial outline of the edge of the middle and outer membrane in IVUS images in advance is omitted, and dividing efficiency is improved.

The invention discloses a captive test (CT) image partitioning method based on adaptive learning. The CT image partitioning method comprises the following steps of: 1) acquiring a CT image; 2) extracting characteristics of the CT image; 3) inputting strokes for representing a lesion region and a non-lesion region on the CT image by a user; 4) constructing a region model of the image by taking the strokes input by the user as a basis according to the extracted characteristics of the CT image, and constructing an edge model of the image by adopting an edge detection method; and 5) combining the region model and the edge model to construct a new model, calculating the new model to obtain a partitioning result. By adopting the method, a difference between the lesion region and the non-lesion region on the CT image can be effectively described; the method adapts to the complexity of the CT image; the problems caused by low signal-to-noise ratio (high noise) of the CT image are solved; the user can quickly and precisely partition the lesion region on the CT image in an interactive manner with high efficiency; and therefore, the production efficiency of a medical department can be greatly improved.

The invention relates to a fast noise-containing image two-dimensional maximum between-class variance threshold value method, which comprises the steps of firstly solving a gray average value and a gray standard deviation of a noise image; smoothing each pixel of the image by adopting an average gray value of a 3*3 neighborhood to acquire a smooth image; then calculating the between-class variance of the smooth image by using a maximum between-class variance threshold value method, reducing the search space of a solution of the between-class variance through the gray average value and the standard deviation, traversing the search space, and recording a solution, which enables the between-class variance to be the maximum, to be an optimal one-dimensional threshold value T0; and calculating a trace of a between-class variance dispersion matrix of a target class and a background class by using a two-dimensional maximum between-class variance method, reducing the search space of a solution of the trace through the optimal one-dimensional threshold value T0 and the gray standard deviation of the noise image, traversing the search space of the solution, and recording a gray value binary group, which enables the trace of the dispersion matrix to be the maximum, to be an optimal two-dimensional cutting threshold value. The method provided by the invention can avoid traversal for all gray levels, and also can acquire an accurate solution while greatly reducing the calculation amount.

The invention discloses an artery blood vessel image model training method, a segmentation method, a device and an electronic device, belonging to the technical field of digital image processing. Theartery blood vessel image model training method comprises the following steps: 1, pre-processing the acquired DSA image to construct an artery blood vessel image database; 2, labeling part of the sample images in the arterial blood vessel image library to construct a labeled sample image set; 3, constructing a convolution depth network and setting parameter of that depth network to generate an initial artery blood vessel segmentation model; 4, training an initial artery blood vessel segmentation model by using a label sample image set to generate an artery blood vessel image segmentation model; 5, further labeling the blood vessel target image obtained by using the artery blood vessel image segmentation model to segment other images except part of the sample images in the artery blood vessel image library, so as to carry out iterative training on the artery blood vessel image segmentation model. The embodiment of the invention can extract target blood vessels from DSA images with highaccuracy.

The invention disclose a method for performing image segmentation by using manifold spectral clustering, which is used for solving the problems of large storage capacity and low computing efficiency and segmentation accuracy in the existing method. The method for performing the image segmentation by using the manifold spectral clustering comprises the following steps: (1) inputting an image, extracting colors and textural features of the image, and obtaining a manifold set of the input image by using a watershed algorithm; (2) computing the manifold feature set, constructing a distance matrix, and acquiring a manifold distance matrix by using the Floyd algorithm; (3) computing a similarity matrix so as to construct a degree matrix and a normalization laplacian matrix; (4) carrying out eigen-decomposition on the normalization laplacian matrix so as to construct a spectral matrix; and (5) normalizing the spectral matrix to obtain a normalization spectral matrix, acquiring the label vector of the manifold set by a K-means algorithm, and outputting a segmentation result. The method for performing the image segmentation by using the manifold spectral clustering has the advantages of small storage capacity and high computing efficiency and segmentation accuracy, and can be used for detecting focal areas of medical images, detecting defects on precision component surfaces, and processing geographic and geomorphic pictures shot by satellites.

The invention discloses a road scene segmentation method based on a residual network and expanded convolution. The method comprises: a convolutional neural network being constructed in a training stage, and a hidden layer of the convolutional neural network being composed of ten Respondial blocks which are arranged in sequence; inputting each original road scene image in the training set into a convolutional neural network for training to obtain 12 semantic segmentation prediction images corresponding to each original road scene image; calculating a loss function value between a set formed by12 semantic segmentation prediction images corresponding to each original road scene image and a set formed by 12 independent thermal coding images processed by a corresponding real semantic segmentation image to obtain an optimal weight vector of the convolutional neural network classification training model. In the test stage, prediction is carried out by utilizing the optimal weight vector of the convolutional neural network classification training model, and a predicted semantic segmentation image corresponding to the road scene image to be subjected to semantic segmentation is obtained. The method has the advantages of low calculation complexity, high segmentation efficiency, high segmentation precision and good robustness.

The embodiment of the invention provides a logic design segmentation method and system, and belongs to the technical field of logic array prototype system verification, and the method specifically comprises the steps: collecting an RTL design file for describing a logic circuit; performing grammatical analysis processing on the RTL design file, extracting an always object and an assign object in the logic model object, respectively packaging, constructing and generating a hypergraph data structure, performing attribute analysis, processing according to the clock domain information to obtain operation frequency information, and performing associated storage on the clock domain information domain and the operation frequency information and corresponding nodes; and performing segmentation processing to obtain corresponding grouped data. By means of the processing scheme, other processing at the back end of the process is not affected, the segmentation duration is shortened, the segmentation efficiency is improved, meanwhile, efficient, reasonable and correct segmentation processing is conducted on the chip design logic content, the performance and efficiency of design segmentation aregreatly improved, then the process of user front-end function verification is accelerated, and the appearance of integrated circuit products is accelerated.

The invention discloses an intelligent Internet-of-Things blind guide stick. A vibration module and a Bluetooth headset module are arranged in a hand-shaped stick handle. A radar ranging module, a GPS positioning module, an LED illumination module, an image recognition module, a bar code recognition module and a cloud data storage module are arranged in a stick body. A charging interface is installed at the bottom of the stick body. The vibration module comprises a microprocessor, a vibration class memory, a vibration time memory and a motor. The Bluetooth headset module comprises a Bluetooth chip, an audio processing chip and an audio emitting module. A voice recognition module and an image recognition module are arranged in the vibration module. Through the intelligent Internet-of-Things blind guide stick, exchange between road information and a user, recognition of the voice of the user, making of a walking route and recognition of blind sidewalk and bus information are realized; whether an obstacle exists in front or not is sensed through ultrasonic waves, meanwhile, a prompter emits sounds to notify the blind, the stick can adapt to people with different heights, and a lamp can caution surrounding pedestrians and vehicles for avoidance; the stick is simple in structure and convenient to use.