69 results about "Target lesion" patented technology

The invention discloses a breast surgery puncture positioning system, which relates to the technical field of image positioning, and is characterized in that a preoperative medical image and an intraoperative ultrasonic image of a patient are acquired to generate a three-dimensional model, an optimal puncture path is planned by receiving a needle insertion point of a user, and the spatial pose of a puncture needle is tracked; the optimal puncture path and the position and posture of the puncture needle are displayed on the real-time ultrasonic image in an overlapped mode, real-time deviation is calculated, the puncture needle is guided and adjusted to the target focus, and real-time compensation is carried out based on the respiratory signal and / or the displacement of the reference datum. Through the multi-modal image fusion and real-time dynamic tracking technology, the puncture precision and hit rate are remarkably improved, meanwhile, the system intelligently plans a safe path and provides an AR navigation interface, the operation difficulty and dependence on doctor experience are greatly reduced, physiological displacement interference is overcome through respiratory gating and a motion compensation mechanism, and the system has the advantages of being high in accuracy and high in reliability. Accuracy and reliability in the whole operation process are guaranteed, and finally standardized and safe mammary gland puncture diagnosis and treatment operation is achieved.

Methods and systems for identifying a region of interest in breast tissue utilize artificial intelligence to confirm that a target lesion identified during imaging the breast tissue using a first imaging modality (e.g. x-ray imaging) has been identified using a second imaging modality (e.g. ultrasound imaging). A computing system operating a lesion matching engine utilizes a machine learning classifier algorithm trained on cases of x-ray images and corresponding ultrasound images in which lesions were identified for further analysis. The lesion matching engine analyzes a target lesion identified with x-ray imaging and a potential lesion identified with ultrasound imaging to determine a likelihood that the target lesion is the same as the potential lesion. A confidence level indicator for the lesion match is presented on a display of a computing device to aid a healthcare provider in locating a lesion in breast tissue.

The invention provides a target focus identification method, system and device based on a medical image, a computer storage medium and a computer program product, and the method comprises the steps: generating an MPR feature vector and a DICOM metadata feature vector which are aligned in dimension based on medical image coding, the MPR feature vector comprises the form and spatial position information of an anatomical structure, and the DICOM metadata feature vector comprises the form and spatial position information of the anatomical structure; the DICOM metadata feature vector comprises scanning parameters, equipment information and patient body position information; the MPR feature vector and the DICOM metadata feature vector are fused through a bidirectional cross attention mechanism, and a comprehensive feature vector containing lesion semantics and space details is generated; and processing the comprehensive feature vector by using a neural network model based on a Transform architecture, and outputting an identification result of the target focus. By adopting the scheme provided by the invention, the complementation of semantic information such as anatomical morphological characteristics and scanning parameters is realized, and the recognition precision of the target focus is improved.

The present invention relates to an image navigation method and system for a surgical microscope. The method includes: first, after fixing a surgical object and markers, radiographing the surgical object and markers to obtain a three-dimensional digital image; then, positioning the markers in the three-dimensional digital image to form a three-dimensional coordinate system and obtain the relative position between the target lesion and the markers; then, plotting a surgical instrument entry point and path guidance points on the three-dimensional digital image to form a navigation path, and transmitting the navigation path data to a media signal processing device; then, positioning the target lesion and markers within the field of view of the surgical microscope and transmitting the image within the field of view of the surgical microscope to the media signal processing device; the media signal processing device recognizing the markers to form a three-dimensional space corresponding to the image within the field of view of the surgical microscope; and real-time navigation is achieved by superimposing the navigation path image at predetermined coordinates on the image within the field of view of the surgical microscope in the three-dimensional space. The system includes a surgical microscope, a media signal processing device, and markers, and the surgical microscope includes a microscope tube and an image enhancement device. By accurately positioning the surgical instrument entry point and path guidance point, the present invention can guide a doctor or related personnel in completing dental treatment and surgery, thereby improving the precision of surgery.

According to the ultrasonic imaging equipment and the lesion distribution presentation method provided by the invention, the volume data is firstly obtained, then the target tissue and the target lesion are identified from the volume data to obtain the three-dimensional contours of the target tissue and the target lesion, and then the three-dimensional schematic diagram is generated based on the two three-dimensional contours. The three-dimensional schematic diagram is used for presenting the form of the target tissue and the form and the position of the target focus in the target tissue, so that a user can intuitively know the distribution condition of the focus in space, the distribution of the focus is easier to understand and master compared with an ultrasonic image, and the working efficiency of a doctor is improved. And generating a two-dimensional diagram corresponding to at least one target section according to the three-dimensional contour of the target tissue and the focus, wherein the two-dimensional diagram corresponding to the target section comprises a target tissue graph used for presenting the form of the target tissue and a target focus graph used for presenting the form and the position of the target focus. Therefore, the user can know the distribution of the focus through the two-dimensional schematic diagram.

The application discloses a self-expanding balloon catheter and relates to the technical field of medical devices, which comprises a main pipe body, a self-expanding balloon and an outer sheath pipe, the outer sheath pipe is sleeved on the main pipe body and is in sliding connection with the main pipe body, the self-expanding balloon is provided with two groups and is located at the proximal end and the distal end of the main pipe body, the main pipe body comprises a medicine injection channel with a medicine release side hole, and the side hole is located between the two groups of self-expanding balloons, and the outer sheath pipe is used for restraining the self-expanding balloon. Through the arrangement, the self-expanding balloon is adopted, the balloon can be self-inflated and expanded, and it is not necessary to rely on external injection of gas or liquid, which is favorable for simplifying the operation process and shortening the operation time; in a non-working state, the balloon is restrained by the outer sheath pipe and is in a contracted state, so that the overall diameter of the catheter is reduced, and it is favorable for sending the catheter to a target lesion position; the outer sheath pipe is withdrawn, the restriction of the balloon is released, the double balloons are self-inflated in a cavity and form a closed medicine perfusion treatment cavity, and it is favorable for realizing local and accurate treatment of non-blood vessel cavity lesions.

The application discloses a kind of indoor light control method, system, equipment and medium, the control method includes: determining the target lesion area of patient;According to the angle information between target lesion area and lighting device, the lighting weight information corresponding to lighting device is calculated;Control lighting device emits light in line with corresponding lighting weight information to target lesion area.The application controls lighting device to emit light according to the lighting weight information calculated according to the angle information between target lesion area and lighting device, can automatically adjust the power or brightness of each lighting device, avoids user manually adjusting the lighting parameter of each lighting device in operating room, solves the problem that composite operating room relies on artificial adjustment shadowless lamp;It improves the intelligent degree of operating room shadowless lamp;Enhance the experience of user.

The invention discloses a sample anti-pollution sampler for gynecological detection, which relates to the technical field of sampling devices, and comprises a mounting part, a grip and a stretching tube, the air exhaust sliding pipe is arranged in the extending pipe in a penetrating manner; the driving mechanism is arranged on the mounting part and used for driving the air exhaust sliding pipe to move and driving the air exhaust sliding pipe to rotate around the axis of the air exhaust sliding pipe in the moving process; the cutting executing mechanism is connected with the stretching-in end of the air exhaust sliding pipe, contracts in the radial direction along with the axial movement of the air exhaust sliding pipe and rotates along with the rotation of the air exhaust sliding pipe so as to conduct girdling on the target tissue; and the sample absorbing mechanism is communicated with an internal channel of the air exhaust sliding pipe, is used for absorbing and accommodating target tissues under the action of negative pressure, and moves along with the axial movement of the air exhaust sliding pipe. According to the device, negative pressure is generated through external air exhaust equipment, target pathological tissues are adsorbed and pulled into the sample adsorption mechanism, and the problems that the pathological tissues do not protrude out of the surface of normal tissues, and positioning and clamping are difficult in a traditional clamping mode are effectively solved.

The present disclosure relates to an oral cavity detection method, device and equipment, comprising: obtaining a three-dimensional tooth model by scanning; detecting a target lesion area of the tooth and its confidence level based on an oral cavity detection model under a current detection view angle; back-projecting the target lesion area to the three-dimensional tooth model to obtain a current lesion position of the three-dimensional tooth model; judging whether the area overlap degree of the current lesion position and a historical lesion position is higher than a preset threshold; if yes, selecting the one with high confidence level as the target lesion position from the current lesion position and the historical lesion position, and updating the mark of the target lesion position; if no, determining the current lesion position as a new lesion position, and adding a mark of the new lesion position; determining a next detection view angle and performing the above operation again until the operation under the last detection view angle is completed, and obtaining multiple lesion positions of the three-dimensional tooth model. The present disclosure can improve the oral cavity detection efficiency and lesion marking accuracy.

The invention discloses a medical aid decision-making method and system based on multi-modal deep learning, and relates to the technical field of auxiliary diagnosis, and the method comprises the steps: obtaining medical image data and non-image data of a target object; based on the medical image data and the non-image data, identifying and segmenting a target focus area, and respectively extracting a depth image feature and a depth non-image feature of the target focus area; the depth image features and the depth non-image features are fused, and comprehensive feature representation of the focus is generated; and segmenting at least one key anatomical structure around the target lesion area by taking the target lesion area as a center. According to the method, the focus-key anatomical structure heterogeneous graph is constructed, and the graph neural network is used for reasoning, so that key pathological interaction relationships such as traction of a nodule to a blood vessel and infringement of the nodule to a bronchial tube can be automatically and quantitatively modeled. Accurate modeling and deep reasoning of the focus-microenvironment complex interaction relation are achieved, and the biological basis of diagnosis and the credibility of the model are improved.

The invention discloses a cone-beam CT arterial phase automatic scanning method and system based on forward projection, and relates to the technical field of medical image.The method comprises the steps that low-dose cone-beam CT pre-scanning is conducted on a target area when a contrast agent is not injected, a three-dimensional image is reconstructed, a to-be-monitored anatomical structure is selected on the three-dimensional image to form a three-dimensional area of interest, and the three-dimensional area of interest is obtained; the method comprises the following steps of: performing forward projection on a three-dimensional region of interest based on cone beam CT geometric parameters to obtain a two-dimensional monitoring region template corresponding to each rack angle, acquiring a real-time perspective image sequence in a contrast agent injection process, calling the corresponding two-dimensional monitoring region template for each frame of image to calculate a monitoring signal, acquiring a baseline value before contrast, calculating a difference value, and acquiring a real-time perspective image sequence; when the difference value reaches a preset threshold value, a trigger signal is output to start arterial phase cone beam CT scanning. Through three-dimensional pre-scanning targeted forward projection and real-time threshold monitoring, the arrival of a target focus contrast agent is accurately judged, the accuracy of the imaging opportunity in the arterial phase is improved, and the radiation dose is reduced.

The invention discloses a PCI drug balloon postoperative effect prediction method, device and equipment based on coronary DSA and plaque property judgment and a storage medium. The PCI drug balloon postoperative effect prediction method comprises the steps that firstly, a PCI preoperative DSA image sequence of a patient to be evaluated is obtained; reconstructing a three-dimensional blood vessel center line of the target lesion segment based on the DSA image sequence, calculating lumen geometric features, inputting the lumen geometric features into a pre-trained plaque property judgment model, and outputting a plaque property judgment result of the target lesion; further combining a plaque property judgment result, DSA-derived vascular morphological parameters, clinical information of a patient and drug balloon instrument parameters to construct a comprehensive feature set; and finally, inputting the comprehensive feature set into a pre-trained regression prediction model, and outputting a PCI drug balloon postoperative prediction effect. According to the method, the plaque type is inversed through the geometrical characteristics of the lumen, the defect that conventional DSA evaluation lacks plaque histological information is overcome, and the reliability of drug balloon postoperative effect prediction is improved.

The present disclosure provides a CT image processing method, device, equipment, storage medium and program product, relating to deep learning technology, AI medical technology, comprising: determining a lung parenchyma region in a computed tomography (CT) image, and cutting out an initial lung parenchyma image from the CT image according to the lung parenchyma region; processing the initial lung parenchyma image to obtain a lesion segmentation mask; processing the initial lung parenchyma image to obtain a predicted lesion region in the initial lung parenchyma image; determining a target lesion region and a target lesion mask corresponding to the target lesion region according to the lesion segmentation mask and the predicted lesion region, wherein the target lesion mask is located in the target lesion region. In this implementation, the lesion region is identified by segmentation and detection, and the target lesion region and the target lesion mask inside the target lesion region are obtained by fusing the two identification results, thereby reducing the probability of missed detection.

A method for measuring a size change of a target lesion in an X-ray image is provided, including receiving a first X-ray image including the target lesion and a second X-ray image including the target lesion, calculating an occupancy of a region corresponding to the target lesion in criterion regions in each of the first X-ray image and the second X-ray image, and measuring a size change of the target lesion based on the calculated occupancies.

A guide extension catheter is provided that can provide a low-friction, large-diameter (for a given compatible guide catheter size) passageway proximal to and optionally through a target lesion within the vasculature, thereby reducing micro- and macro-vasculature damage due to the delivery of an interventional device. A guide extension catheter is disclosed that is positionable within a guide catheter and configured to receive an interventional device for insertion into the vascular system. The guide extension catheter includes a push member, a first reinforcing member in contact with the push member, and a radially collapsible tubular membrane in contact with the push member and the first reinforcing member. The tubular membrane is positioned distal to the first reinforcing member and is radially collapsed or wrapped around the push member prior to receiving the interventional device. The push member provides sufficient column strength to prevent longitudinal collapse of the tubular membrane during use.

The invention discloses a vascular intervention catheter integrated with a real-time imaging function and an interventional therapy system.The vascular intervention catheter integrated with the real-time imaging function comprises a catheter body and an imaging probe, a working channel is formed in the catheter body, the imaging probe is arranged on the side wall of the far end of the catheter body, and in the process that the catheter body stretches into a blood vessel, the working channel is communicated with the imaging probe. The imaging probe can acquire image information in a blood vessel in real time, the catheter body stops moving when stretching into a target focus position, then corresponding surgical operation can be carried out, and the working channel is used for the working part to carry out operation according to the image information acquired by the imaging probe, so that the function of operating while observing is achieved. By means of the vascular intervention catheter, continuous visualization of the whole operation process can be achieved; the first-time passing success rate and the overall efficiency of the operation are improved; and the safety and controllability of the operation are improved.

The present application relates to the technical field of medical devices, and provides a continuum mechanical arm structure integrated with laser focusing and ultrasonic feedback functions, comprising a distal continuum module, a proximal continuum module, a first driving tendon and a second driving tendon, the first driving tendon being used for controlling movement of the distal continuum module, and the second driving tendon being used for controlling movement of the proximal continuum module; a pose detection module and a laser optical module are sequentially arranged at the front end of the distal continuum module, the pose detection module is used for acquiring relative pose information of a target lesion, and the laser optical module is used for projecting laser energy to treat the target lesion; an endoscope vision module is arranged on the side wall of the distal continuum module, and the endoscope vision module is used for acquiring image information. The present application not only can improve the flexibility and adaptability of surgery, but also can realize precise treatment, improve treatment effect, avoid damage to surrounding healthy tissues, reduce surgical risk, and meet the requirements of flexibility and accuracy of orthopedic treatment.

The invention provides a target detection method and device, equipment and a storage medium, which are used for solving the problem of low target detection accuracy. The method at least comprises the following steps: based on biological tissue characteristics extracted from a to-be-detected image, carrying out region division on a region where abnormal tissues are located in the to-be-detected image to obtain at least one lesion reference region; based on the visual pathological features extracted from the to-be-detected image, predicting lesion positions in the to-be-detected image, and obtaining a plurality of candidate lesion positions; aiming at the at least one lesion reference area, respectively executing the following steps: selecting a target lesion position of which the position relationship meets a position association condition from the plurality of candidate lesion positions based on the position relationship between each of the plurality of candidate lesion positions and one lesion reference area; and obtaining a target detection result based on the obtained at least one target lesion position. And at least one target lesion position with the most suspected lesion and the most representative is screened out through the position relationship, so that the target detection accuracy is improved.

The application provides an insertion part and an endoscope, and relates to the technical field of endoscopes, and solves the problem of large displacement of the insertion part relative to a target lesion caused by breathing movement in the prior art. The endoscope is provided with an insertion part, a working channel, a first fluid channel and a second fluid channel which are independently arranged in the front end assembly of the insertion part and extend from the proximal end to the distal end; the outer peripheral wall of the insertion part has a to-be-fixed region which is adjacent to the front end assembly as a whole; the first fluid channel delivers a hydrogel to an outlet located in the to-be-fixed region for discharge; the second fluid channel delivers a hydrogel degrading agent to an outlet located in the to-be-fixed region for discharge; and the outlet of the working channel is located on the distal end face of the insertion part for passing out of an instrument and negative pressure suction. The application can fill the insertion part with a hydrogel between the insertion part and the bronchus through the first fluid channel to temporarily fix the insertion part at the position adjacent to the lesion in the bronchus, so that the displacement degree between the insertion part and the lesion can be reduced during the breathing process of the patient.

The application discloses an image processing method based on artificial intelligence and belongs to the technical field of image processing. The application sets a sliding window to an ultrasonic image, acquires row and column texture correlation of pixel points, converts and exponentially amplifies irrelevance, filters row and column lesion points, extracts the same target lesion points, constitutes first and second lesion target regions, cuts the region boundary, acquires sharpness and branching degree and constitutes a sequence, processes the sequence through a lesion region recognition neural network, and obtains lesion region classification. The application improves classification accuracy through accurate extraction of lesion texture and morphological features and intelligent analysis, and provides an efficient technical scheme for ultrasonic lesion diagnosis.

The invention discloses a multistage drug-loaded nano motor and a preparation method and application thereof, the nano motor is based on a magnetic nanoparticle core (Fe3O4), the surface of the nano motor is modified with a mesoporous silica layer of a sandwich structure to load thrombolytic drugs, and the outermost layer is wrapped with polydopamine. The Fe3O4 core enables the nano-particle to target a lesion area through physical adsorption, the sandwich structure further improves the dispersibility and uniformity of the nano-particle, and the polydopamine shell can play photo-thermal thrombolysis, anti-inflammatory and anti-oxidation roles under 808nm laser irradiation, so that systematic treatment for diseases is realized. The preparation method provided by the invention has the advantages of simplicity in operation, high universality, convenience in large-scale production and the like, and the prepared nanoparticles can overcome blood flow resistance, target a focus area and play a comprehensive treatment function, and have potential application prospects and practical significance in the biomedical fields of tissue repair, focus ablation and the like.

The invention discloses an eye medical data collection system and method based on artificial intelligence, and relates to the technical field of eye medical data collection, and the method comprises the steps: S10, generating an eye lesion feature data set of a to-be-detected eye of a target patient; s20, acquiring a first eye lesion feature data set and a second eye lesion feature data set based on a discrete condition of a target transaction deviation angle in the target lesion feature data set; s30, analyzing and determining an auxiliary eye lesion feature data set of the to-be-detected eyes of the target patient; and S40, generating eye medical data of the to-be-tested eyes of the target patient. The abnormal lesion feature data can be visually and clearly identified through the data set, the analysis efficiency of the data set is improved to a certain extent, and the generated eye medical data is more representative and less in data size compared with eye medical data directly acquired through an eye detection instrument.

The invention discloses a device for cerebral hemorrhage risk assessment in the perioperative period of a cardiovascular surgery and a storage medium. The apparatus comprises: a processor; the device realizes the following operations: extracting risk key features, an initial risk assessment result and an SHAP value corresponding to each risk key feature; screening a target lesion type, and calling a preset segmentation model to perform lesion segmentation to obtain a lesion segmentation result; based on the image data and the focus segmentation result, calculating a dynamic weight and endowing a coding feature to obtain a weighted image feature; constructing a knowledge graph; extracting a target sub-image set according to clinical information and an entity set corresponding to the focus segmentation result; obtaining text features; constructing a cross-feature collaborative calibration matrix for feature calibration; and inputting the calibrated features into a multi-mode decoder for decoding to obtain a final risk assessment result. By using the scheme of the invention, accurate and credible risk assessment results and decision support can be provided for clinic.

The application discloses a lesion recognition model training method and device, equipment and a storage medium. The method comprises the following steps: obtaining model training information for guiding model training. According to the model training information, training sample data for model training is obtained. According to a preset lesion recognition model and the training sample data, a training task corresponding to the sample label is executed. When the number of training rounds of the preset lesion recognition model meets a preset round threshold, it is determined that the training of the preset lesion recognition model is completed, and a target lesion recognition model corresponding to the model training information is obtained. The technical scheme can automatically execute the corresponding training task according to the training sample label, enhance the personalized model training of the complex analysis scene, and promote the customized training and development and application of the medical model through the code-free configuration mode.

The utility model discloses an anti-condensation precise drug release pipeline device for an upper digestive tract, which comprises a triple-combination drug delivery pipe used for maintaining drug particles atomized by drug atomization equipment in a micron-sized state all the time and delivering the drug particles to a target focus area, the three-layer structure of the triple-combination medicine feeding pipe sequentially comprises a hydrophobic structure layer, a supporting structure layer and a heat insulation structure layer from inside to outside, and the polytetrafluoroethylene layer serves as the hydrophobic structure layer, has extremely high hydrophobicity and prevents medicine particles from being condensed or deposited in the pipe; the glass fiber layer serves as a supporting structure layer and wraps the polytetrafluoroethylene layer, and it is guaranteed that the pipeline is kept stable; the polyurethane foam layer is used as a heat insulation structure layer, wraps the glass fiber layer, has excellent biocompatibility and heat insulation effect, and avoids the influence of temperature fluctuation on the drug stability; according to the triple-combination medicine delivery tube, medicine can be efficiently and accurately delivered to a target focus area, and therefore the treatment effect is improved.

Devices, systems, and methods used to deploy a localization device into a target lesion are disclosed. The system includes a delivery device configurable in a ready-to-deploy state and a deployed state. When the delivery device is in the ready-to-deploy state, biased engagement members are engaged with proximal detents of a handle of the delivery device. When the delivery device is in the deployed state, the biased engagement members are engaged with distal detents. The delivery device is transitioned from the ready-to-deploy state to the deployed state by distal displacement of a plunger relative to the handle. The localization device is deployed from a cannula of the delivery device into the target lesion when the delivery device transitions from the ready-to-deploy state to the deployed state.

The invention relates to a lesion assessment method and device. The method comprises the following steps: acquiring a current report text and a historical report text corresponding to a medical image obtained by scanning a patient twice; utilizing a preset focus analysis model to extract a focus described by the current image in the current report text, and determining structured information of the focus according to a description described by the historical image in the historical report text; and inputting the structural information of the focus and the historical examination conclusion description in the historical report text into a preset focus evaluation model, determining a target focus, evaluating the target focus according to a preset reference evaluation rule, and outputting an evaluation result of the target focus. The method can improve the accuracy of the lesion evaluation result.

The application provides a processing method and device for follow-up case data, equipment and a storage medium. The processing method comprises: determining a first detection result of a target lesion in a first space according to position information and size information of the target lesion in a first medical image; performing transformation processing on the first detection result according to a registration transformation matrix between the first medical image and a second medical image to obtain an initial lesion matching result of the first detection result in a second space; and inputting the first medical image, the second medical image, the first detection result and the initial lesion matching result into a feature extraction model to output a lesion feature extraction result of the second medical image. In this way, the application enables the model to effectively combine the anatomical structure information of the lesion on the basis of the medical image information, thereby improving the feature extraction accuracy of the model for the same lesion in different medical images and the precision of matching and positioning of the same lesion in follow-up case data.

The application discloses an X-ray guided puncture positioning method and system based on infrared optical coordinates, and belongs to the technical field of puncture positioning, and comprises the following steps: an operator forms a body surface infrared coordinate system on the body surface of a patient by using an infrared coordinate generating device; the operator operates an X-ray machine to perform perspective and obtain a real-time X-ray image coordinate system containing a target lesion, and establishes a space mapping relationship between the body surface infrared coordinate system and the real-time X-ray image coordinate system; after the virtual intersection point of the body surface infrared coordinate line coincides with the target lesion point on the image, the actual physical intersection point on the body surface is determined as the best body surface puncture needle entry point and is marked; and the operator performs a puncture operation according to the direction of the body surface infrared coordinate line, the best body surface puncture needle entry point determined above and the puncture angle provided above. The application solves the problem that the prior art lacks an auxiliary scheme capable of realizing fast, accurate and low-cost body surface positioning without interfering with an X-ray image.

The present application provides a kind of lightweight medical image segmentation method, device, medium and electronic equipment, utilize the target lesion in the medical image to be segmented in pre-trained lightweight medical image segmentation model and carry out segmentation, in lightweight medical image segmentation model, down-sampling unit carries out down-sampling processing to medical image to be segmented;First EERM layer carries out feature extraction to the down-sampling processing result and obtains preliminary feature extraction result;First segmentation unit carries out down-sampling processing to preliminary feature extraction result and then carries out feature extraction operation based on ConvLSTM;Second segmentation unit carries out feature extraction operation based on ConvLSTM to preliminary feature extraction result;Third segmentation unit is used to carry out feature extraction operation based on ConvLSTM to preliminary feature extraction result after carrying out up-sampling processing;Result output unit fuses three feature extraction results and then outputs image segmentation result.The present application can run in the device with relatively lower graphics card configuration, model parameter is less, and can give consideration to calculation speed and calculation accuracy.