Method for recognition and reduction of blood speckle in blood vessel imaging system

Claim 1

A method for enhancing an intravascular ultrasonic blood vessel image comprising the steps of: a) illuminating an intravascular target from within a blood vessel with ultrasonic energy to generate ultrasonic echoes from said intravascular target; thereafter b) intercepting the ultrasonic echoes from the intravascular target; c) transforming the ultrasonic echoes into electrical signals; d) forming and constructing an image based on amplitude of each said electrical signals, by illuminating and sweeping said intravascular target and by circumferentially and radially sampling the electrical signals, assigning an initial gray scale value to each increment (pixel) in said image; e) computing a distance metric for each pixel using seed values for each said pixel; f) sampling said image to obtain a sampled image; g) growing a first interior classification for points in said a sampled image from a first seed pixel preselected to be entirely within a first region of a first known class; h) growing a second exterior classification for exterior points from a second seed pixel in a second region of a second known class; i) identifying and classifying other pixels in the a sampled image by searching for points between interior classified pixels and exterior classified pixels and tagging said other pixels as said edge pixels; j) remapping all interior pixels, exterior pixels and edge pixels so classified into a full resolution image 1) by mapping exterior pixels with first gray scale values, 2) by mapping interior pixels with second gray scale values and 3) by mapping edge pixels according to a high resolution Fisher classification for said edge pixels; and k) outputting said full resolution image for display.

Claim 2

A method for enhancing an intravascular ultrasonic blood vessel image comprising the steps of: a) illuminating an intravascular target from within a blood vessel with ultrasonic energy to generate ultrasonic echoes from said intravascular target; thereafter b) intercepting the ultrasonic echoes from the intravascular target; c) transforming the ultrasonic echoes into electrical signals; d) forming and constructing an image based on amplitude of each said electrical signals, by illuminating and sweeping said intravascular target and by circumferentially and radially sampling the electrical signals, assigning an initial gray scale value to each increment (pixel) in said image; e) computing a distance metric for each pixel using seed values for each said pixel; f) sampling said image to obtain a sampled image; g) growing a first interior classification for points in said a sampled image from a first seed pixel preselected to be entirely within a first region of a first known class; h) growing a second exterior classification for exterior points from a second seed pixel in a second region of a second known class; i) identifying and classifying other pixels in the a sampled image by searching for points between interior classified pixels and exterior classified pixels and tagging said other pixels as said edge pixels; j) remapping all interior pixels, exterior pixels and edge pixels so classified into a full resolution image 1) by mapping exterior pixels with first gray scale values, 2) by mapping interior pixels with second gray scale values and 3) by mapping edge pixels according to a high resolution Fisher classification for said edge pixels; and k) outputting said full resolution image for display.