Cardiac view recognition and disease recognition

摘要

A method for recognizing heart diseases in a cardiac echo video of a heart with an unknown disease using a spatio-temporal disease model derived from a training echo video, comprising the steps of: generating a plurality of training models for heart diseases, wherein the cardiac echo videos are each derived from a known viewpoint and the disease of the heart is known; analyzing the video of the heart with the unknown disease by fitting a model of shape and motion for each frame and combining the results across the frames; and, reporting the disease using a classification method for choosing among the diseases of interest.

主权项

1. A method for determining a transducer position and a transducer viewpoint using spatial and temporal information in a cardiac echo video, comprising:
receiving a spatial and temporal model for a known transducer viewpoint from a sample learning data; analyzing a new single heart-cycle echo sequence for the known transducer viewpoint by fitting a spatial aspect of the spatial and temporal model to each frame in an annotated cardiac echo cycle video, thereby creating a time-varying set of spatial features; fitting the set of spatial features to a motion model of the cardiac echo cycle, wherein the motion model is derived from the annotated cardiac echo cycle video; evaluating the spatial and temporal model fit using a combined fit of the motion model; receiving an appearance of a plurality of heart-cycle variations using the plurality of heart-cycle variations for isolated features via the spatial and temporal models from the sample learning set; and determining a matching model using a matching algorithm for recognizing a cardiac echo view from the appearance of the plurality of heart-cycle variations; wherein the matching algorithm estimates a measure of a sequence fit for each cardiac view, in which the sequence fit is an average appearance fit over the cardiac echo cycle, a Mahalannobis distance on motion, and a motion reconstruction error term.