Abstract: Disclosed are a coronary artery segmentation method, an electronic device, and a computer-readable storage medium. The coronary artery segmentation method includes: iteratively performing a detection step for a to-be-detected image; for a plurality of the detection steps corresponding to the to-be-detected image, selecting the to-be-detected region in a current detection step from the replaced to-be-detected image in a previous detection step, and using a part of the comparison result in the previous detection step as the regional coronary artery inspiration data in the current detection step; and outputting a finally replaced to-be-detected image as coronary artery segmentation data. This method implements precise segmentation, and effectively distinguishes between a coronary artery and a vein false positive.

Abstract: Disclosed are a method, a system, and an apparatus for generating guide information, and an electronic device. The method includes: obtaining image data to be identified of a detection target; segmenting the image data to be identified by using a trained image segmentation model to obtain segmented target image data; determining image data of positioning patches arranged on a body surface of the detection target in the target image data as feature points, and establishing a mapping relationship between a preoperative spatial three-dimensional image coordinate system and an intraoperative spatial positioning coordinate system; and generating guide information based on the mapping relationship, to guide a puncture needle to perform puncturing on the detection target. An operation time of a puncture operation can be greatly shortened, and puncture efficiency and accuracy can be improved.

Abstract: Disclosed are a method and an apparatus of processing image. The method includes: obtaining an initial bone segmentation result; and fusing the initial bone segmentation result based on characteristics of and correspondences between a plurality of bone segmentation results in the initial bone segmentation result, to obtain a target bone segmentation result. The initial bone segmentation result includes the plurality of bone segmentation results generated by a plurality of different deep learning models. Methods in the embodiments of the present application can improve precision of a fusion result of the plurality of bone segmentation results.

Abstract: In an example embodiment, a deep machine learning model ranks cohorts of users as well as cohorts of products in a single ranking. When utilized to determine which cohort members to display to a user, the system selects one user cohort and one product cohort as the “best” (e.g., the top ranked user cohort and the top ranked product cohort). This ranking may be based on a number of contextual and non-contextual features, including viewer features (characteristics of the user operating the user interface), viewee features (characteristics of or related to the litem that the user is viewing, such as the characteristics of another user whose profile the user is viewing), and viewer-viewee relationship features (indications about how the viewer and viewee are related, such as common schools, locations, places of employment, etc.).

Abstract: One aspect of the present disclosure pertains to a method of forming a semiconductor device. The method includes receiving a workpiece having a first metal gate stack over a first channel region, a second metal gate stack over a second channel region, a source/drain (S/D) feature between the first and second channel regions, and an S/D contact over the S/D feature. First and second dielectric caps are formed over the first and second metal gate stacks and a contact etch stop layer (CESL) is formed over the S/D contact and over the first and second dielectric caps. An interlayer dielectric (ILD) layer is formed over the CESL and an S/D via trench is formed through the ILD layer and the CESL. An S/D via is formed in the S/D via trench, making full surface contact with the S/D contact and partial surface contact with the first and second dielectric caps.

Abstract: The method for image segmentation includes: acquiring, according to an image to be segmented including a background, a mediastinum, an artery and a vein, a first segmentation result of the mediastinum, the artery, the vein and the background in a mediastinum region of the image to be segmented; acquiring, according to the image to be segmented, a second segmentation result of a blood vessel and the background in an epitaxial region of the image to be segmented; and acquiring, according to the first segmentation result and the second segmentation result, a segmentation result of the mediastinum, the artery, the vein and the background of the image to be segmented, so that the segmentation accuracy and the segmentation efficiency of the artery and the vein may be improved.

Abstract: Disclosed are an image matching method, an image matching device, and a storage medium. A first image sequence and a second image sequence are acquired, and thus a first object and a second object are reconstructed and generated based on the first image sequence and the second image sequence respectively. The registration of the first object and the second object are further performed, and a mapping relationship obtained according to a registration result may indicate a correspondence between image frames in the first image sequence and image frames in the second image sequence. Compared with setting a difference value artificially, obtaining the correspondence between image frames in the first image sequence and image frames in the second image sequence by using the image matching method improves the matching accuracy.

Abstract: Disclosed is a speech synthesis method including: acquiring fundamental frequency information and acoustic feature information from original speech; generating an impulse train from the fundamental frequency information, and inputting it to a harmonic time-varying filter; inputting the acoustic feature information into a neural network filter estimator to obtain corresponding impulse response information; generating noise signal by a noise generator; determining, by the harmonic time-varying filter, harmonic component information through filtering processing on the impulse train and the impulse response information; determining, by a noise time-varying filter, noise component information based on the impulse response information and the noise; and generating a synthesized speech from the harmonic component information and the noise component information.

Abstract: Disclosed is a speech synthesis method including: acquiring fundamental frequency information and acoustic feature information from original speech; generating an impulse train from the fundamental frequency information, and inputting it to a harmonic time-varying filter; inputting the acoustic feature information into a neural network filter estimator to obtain corresponding impulse response information; generating noise signal by a noise generator; determining, by the harmonic time-varying filter, harmonic component information through filtering processing on the impulse train and the impulse response information; determining, by a noise time-varying filter, noise component information based on the impulse response information and the noise; and generating a synthesized speech from the harmonic component information and the noise component information.

Abstract: Disclosed are a method for image segmentation, and an electronic device. The method includes: performing, based on a preliminary segmentation result of a cerebral hematoma in a head medical image, a secondary segmentation of the cerebral hematoma on a pre-processed head medical image, to obtain a secondary segmentation result of the cerebral hematoma in the pre-processed head medical image; and acquiring a final segmentation result of the cerebral hematoma in the head medical image based on the secondary segmentation result, so that a segmentation effect of a cerebral hematoma in a head medical image can be improved in a case that the head medical image has data heterogeneity.

Abstract: Disclosed are a method, a system, and an apparatus for generating guide information, and an electronic device. The method includes: obtaining image data to be identified of a detection target; segmenting the image data to be identified by using a trained image segmentation model to obtain segmented target image data; determining image data of positioning patches arranged on a body surface of the detection target in the target image data as feature points, and establishing a mapping relationship between a preoperative spatial three-dimensional image coordinate system and an intraoperative spatial positioning coordinate system; and generating guide information based on the mapping relationship, to guide a puncture needle to perform puncturing on the detection target. An operation time of a puncture operation can be greatly shortened, and puncture efficiency and accuracy can be improved.

Abstract: Disclosed are a method and an apparatus of processing image. The method includes: obtaining an initial bone segmentation result; and fusing the initial bone segmentation result based on characteristics of and correspondences between a plurality of bone segmentation results in the initial bone segmentation result, to obtain a target bone segmentation result. The initial bone segmentation result includes the plurality of bone segmentation results generated by a plurality of different deep learning models. Methods in the embodiments of the present application can improve precision of a fusion result of the plurality of bone segmentation results.

Abstract: Disclosed are a segmentation method for a pneumonia sign, a computer-readable storage medium, and an electronic device. A lung region image in a CT image is input into each of a plurality of neural network models to obtain a plurality of pneumonia sign images separately; and the plurality of pneumonia sign images are combined to obtain a pneumonia comprehensive sign image. Pneumonia sign images in a large number of to-be-detected CT images may be obtained efficiently and accurately by using the neural network models, so as to provide reliable data basis for subsequent pneumonia diagnosis.

Abstract: Disclosed are a method and an apparatus for analyzing medical treatment data based on deep learning. The core content of the method is the establishment of a model in a computer using a deep convolution neuron algorithm in deep learning. The model assists the doctors to make correct judgments and effective decisions for the large amount of medical treatment data using mass medical treatment data selection and optimization model parameters by “training” the model to automatically learn a pathology analysis process of the doctors or the medical researchers and then helping them in processing the large amount of medical treatment data. The present invention can greatly reduce the work stress for the doctors or the medical researchers and improve the work efficiency thereof, and can free the doctors or the medical researchers from heavy analysis work on the medical treatment data or medical data.

Abstract: The method for image segmentation includes: acquiring, according to an image to be segmented including a background, a mediastinum, an artery and a vein, a first segmentation result of the mediastinum, the artery, the vein and the background in a mediastinum region of the image to be segmented; acquiring, according to the image to be segmented, a second segmentation result of a blood vessel and the background in an epitaxial region of the image to be segmented; and acquiring, according to the first segmentation result and the second segmentation result, a segmentation result of the mediastinum, the artery, the vein and the background of the image to be segmented, so that the segmentation accuracy and the segmentation efficiency of the artery and the vein may be improved.

Abstract: A method for analysing medical treatment data based on deep learning via an intelligent analyser thereof, effectively alleviates work stress for doctors in a hospital or medical researchers and performs scientific analysis on a large amount of medical treatment data or medical data and obtain an analysis result matched therewith. The core content of the method is the establishment of a model in a computer using a deep convolution neuron algorithm in deep learning. The model assists the doctors to make correct judgments and effective decisions for the large amount of medical treatment data using mass medical treatment data selection and optimization model parameters by “training” the model to automatically learn a pathology analysis process of the doctors or the medical researchers and then helping them in processing the large amount of medical treatment data or medical data.

Abstract: Disclosed are an image matching method, an image matching device, and a storage medium. A first image sequence and a second image sequence are acquired, and thus a first object and a second object are reconstructed and generated based on the first image sequence and the second image sequence respectively. The registration of the first object and the second object are further performed, and a mapping relationship obtained according to a registration result may indicate a correspondence between image frames in the first image sequence and image frames in the second image sequence. Compared with setting a difference value artificially, obtaining the correspondence between image frames in the first image sequence and image frames in the second image sequence by using the image matching method improves the matching accuracy.

Abstract: A method and device for performing based learning on a medical image includes reading raw data of a medical image, performing transformation processing on the data by analyzing a data attribute, and integrating the same into a data format capable of being received by a model to be trained; selecting a transformation method by comparing parameters of the model to be trained and a trained model, so as to perform parameter transformation and apply transformation-based learning to training of the model to be trained for the medical image; and upon finishing model training, applying a parameter of a trained model to image category analysis. The invention further includes a device for performing transformation-based learning on a medical image, including: a data processing module; a transformation-based learning module; and an application module.

Abstract: A computed tomography (CT) pulmonary nodule detection method based on deep learning is provided. The method comprises the steps of: acquiring 3D pulmonary CT sequence images of a user; processing the acquired 3D pulmonary CT sequence images into 2D image data; inputting 2D image data into a preset deep learning network model for training to obtain a trained pulmonary nodule detection model; inputting a set of 3D pulmonary CT sequence images to be tested into the trained pulmonary nodule detection model to obtain a preliminary pulmonary nodule detection result; applying a pulmonary region segmentation algorithm based on deep learning to the preliminary pulmonary nodule detection result to remove false positive pulmonary nodules, so as to obtain a final pulmonary nodule detection result.

Abstract: A computed tomography (CT) pulmonary nodule detection method based on deep learning is provided. The method comprises the steps of: acquiring 3D pulmonary CT sequence images of a user; processing the acquired 3D pulmonary CT sequence images into 2D image data; inputting 2D image data into a preset deep learning network model for training to obtain a trained pulmonary nodule detection model; inputting a set of 3D pulmonary CT sequence images to be tested into the trained pulmonary nodule detection model to obtain a preliminary pulmonary nodule detection result; applying a pulmonary region segmentation algorithm based on deep learning to the preliminary pulmonary nodule detection result to remove false positive pulmonary nodules, so as to obtain a final pulmonary nodule detection result.