Abstract: Various embodiments described herein provide for detection of a particular object within a scene depicted by image data by using a coarse-to-fine approach/strategy based on one or more relationships of objects depicted within the scene.

Abstract: Various embodiments described herein provide for detection of a particular object within a scene depicted by image data by using a coarse-to-fine approach/strategy based on one or more relationships of objects depicted within the scene.

Abstract: Various embodiments described herein provide for a neural network tailored, based on user-provided input data, to detect user-specified objects in image data. An architecture of an embodiment may use unlabeled data from the user, such as a set of images from a video camera stream, while parameters of a tailored neural network (CNN) are trained or adapted.

Abstract: Various embodiments described herein provide for a neural network tailored, based on user-provided input data, to detect user-specified objects in image data. An architecture of an embodiment may use unlabeled data from the user, such as a set of images from a video camera stream, while parameters of a tailored neural network (CNN) are trained or adapted.

Abstract: A method for generating an ultrasound image includes receiving a sequence of intravascular ultrasound (IVUS) data obtained as an IVUS imager moves through a body lumen; identifying at least one bifurcation of the body lumen from the sequence of IVUS data; determining a bifurcation angle between two branches of the body lumen; and displaying a longitudinal view of the body lumen using the IVUS data and incorporating the bifurcation angle to angularly align portions of the longitudinal view corresponding to the two portions of the body lumen.

Abstract: Various embodiments described herein provide for detection of a particular object within a scene depicted by image data by using a coarse-to-fine approach/strategy based on one or more relationships of objects depicted within the scene.

Abstract: Various embodiments described herein provide for detection of a particular object within a scene depicted by image data by using a coarse-to-fine approach/strategy based on one or more relationships of objects depicted within the scene.

Abstract: Various embodiments described herein provide for a neural network tailored, based on user-provided input data, to detect user-specified objects in image data. An architecture of an embodiment may use unlabeled data from the user, such as a set of images from a video camera stream, while parameters of a tailored neural network (CNN) are trained or adapted.

Abstract: A method for imaging a patient using an intravascular ultrasound imaging system includes inserting a catheter into a target imaging region within patient vasculature. The catheter includes at least one transducer coupled to a control module. A plurality of image frames is generated as the at least one transducer is moved within patient vasculature. The image frames are based on data obtained from a plurality of echo signals reflected from patient tissue in response to a plurality of acoustic signals output from the at least one transducer. Each image frame includes a motion-blur component. A motion-blur-intensity estimate is quantified for each image frame to form a motion-blur-estimation signal that varies over time and that includes periodic local motion blur minima. Temporal variations of the motion-blur components are filtered to reduce variation in the motion-blur-intensity estimate arising from vessel structure variation. Image frames are selected that contain local motion blur minima.

Abstract: A method for processing a sequence of ultrasound frames for display includes receiving a sequence of ultrasound frames; characterizing a morphology of regions within a plurality of the ultrasound frames; determining a frame profile for each of the plurality of ultrasound frames; detecting a plurality of key frames from the plurality of ultrasound frames using the frame profiles of the plurality of ultrasound frames; and displaying at least one of the key frames.

Abstract: A method for generating an ultrasound image includes receiving a sequence of intravascular ultrasound (IVUS) data obtained as an IVUS imager moves through a body lumen; identifying at least one bifurcation of the body lumen from the sequence of IVUS data; determining a bifurcation angle between two branches of the body lumen; and displaying a longitudinal view of the body lumen using the IVUS data and incorporating the bifurcation angle to angularly align portions of the longitudinal view corresponding to the two portions of the body lumen.

Abstract: A method for imaging a patient using an intravascular ultrasound imaging system includes inserting a catheter into a target imaging region within patient vasculature. The catheter includes at least one transducer coupled to a control module. A plurality of image frames is generated as the at least one transducer is moved within patient vasculature. The image frames are based on data obtained from a plurality of echo signals reflected from patient tissue in response to a plurality of acoustic signals output from the at least one transducer. Each image frame includes a motion-blur component. A motion-blur-intensity estimate is quantified for each image frame to form a motion-blur-estimation signal that varies over time and that includes periodic local motion blur minima. Temporal variations of the motion-blur components are filtered to reduce variation in the motion-blur-intensity estimate arising from vessel structure variation. Image frames are selected that contain local motion blur minima.