Automatic extraction of disease-specific features from doppler images

1. A method for automatic extraction of disease-specific features from Doppler images, comprising:

obtaining a raw Doppler image from a series of images of an echocardiogram;

isolating a region of interest from the raw Doppler image, the region of interest (i) including a Doppler image and an electrocardiogram (ECG) signal, and (ii) depicting at least one heart cycle;

determining a velocity envelope of the Doppler image in the region of interest;

extracting the electrocardiogram (ECG) signal by detecting the electrocardiogram (ECG) signal using an energy maximization equation:

E ( i,j )=λ 1 E continuity ( i,j )+λ 2 E color ( i,j )+λ 3 E gradient ( i,j )+λ 4 E notgray ( i,j )

wherein E(i,j) is an energy value, E continuity (i,j) is a continuity of the electrocardiogram (ECG) signal, E color (i,j) is a color profile of the electrocardiogram (ECG) signal, E gradient (i,j) is a gradient between the color profile of the electrocardiogram (ECG) signal and a background of the image, E notgray (i,j) is RGB value of the electrocardiogram (ECG) signal that is not gray, and each of, λ 1 , λ 2 , λ 3 , and λ 4 is a weighting factor, and

synchronizing the extracted the electrocardiogram (ECG) signal with the Doppler image over the at least one heart cycle, within the region of interest;

calculating a value of a clinical feature based on the extracted the electrocardiogram (ECG) signal synchronized with the velocity envelope; and

comparing the value of the clinical feature with clinical guidelines associated with the clinical feature to determine a diagnosis of a disease.

2. The method of claim 1 , wherein the obtaining the raw Doppler image includes categorizing the series of images using optical character recognition and machine learning.

3. The method of claim 1 , wherein the determining the velocity envelope includes tracing the Doppler image after applying a foreground/background separation technique.

4. The method of claim 1 , further comprising:

extracting a clinical annotation of the Doppler image when the clinical annotation is present in the echocardiogram, and incorporating the clinical annotation when determining the velocity envelope.

5. The method of claim 1 , wherein the clinical feature is a maximum jet velocity.

6. The method of claim 1 , wherein the clinical feature is a mean pressure gradient.

7. The method of claim 1 , wherein the velocity envelope includes an upper envelop associated with a positive velocity value.

8. The method of claim 1 , wherein the velocity envelope includes a lower envelop associated with a negative velocity value.

9. The method of claim 1 , wherein the raw Doppler image is at least one of a continuous wave (CW) Doppler image and a pulse wave (PW) Doppler image.

10. The method of claim 1 , wherein the disease is aortic stenosis.

11. The method of claim 1 , wherein the obtaining the raw Doppler image includes categorizing the series of images using optical character recognition and machine learning.

12. The method of claim 1 , further comprising:

extracting a clinical annotation of the Doppler image when the clinical annotation is present in the echocardiogram, and incorporating the clinical annotation when determining the velocity envelope to improve an accuracy of the velocity envelope.

13. A method for deriving a clinical feature from Doppler images to diagnose a valvular disease, comprising:

tracing a Doppler image located within a region of interest to extract an upper velocity envelope and a lower velocity envelope of the Doppler image, wherein the region of interest is created from a raw Doppler image, the raw Doppler image being obtained from a series of images of an echocardiogram;

synchronizing an electrocardiogram (ECG) signal with the Doppler image within the region of interest, over at least one heart cycle, the electrocardiogram (ECG) signal being detecting using a maximum energy function:

E ( i,j )=λ 1 E continuity ( i,j )+λ 2 E color ( i,j )+λ 3 E gradient ( i,j )+λ 4 E notgray ( i,j )

wherein E(i,j) is an energy value, E continuity (i,j) is a continuity of the electrocardiogram (ECG) signal, E color (i,j) is a color profile of the electrocardiogram (ECG) signal, E gradient (i,j) is a gradient between the color profile of the electrocardiogram (ECG) signal and a background of the image, E notgray (i,j) is RUB value of the electrocardiogram (ECG) signal that is not gray and each of, λ 1 , λ 2 , λ 3 , and λ 4 is a weighting factor; and

determining at least one of a maximum jet velocity and a mean pressure gradient to diagnose the valvular disease of a patient.

14. The method of claim 13 , wherein the electrocardiogram (ECG) signal is detected using an energy maximization equation.

15. The method of claim 13 , further comprising:

extracting and incorporating a clinical annotation of the Doppler image when the clinical annotation is present in the echocardiogram, to improve the determines velocity envelope.

16. A method for automatic extraction of disease-specific features from Doppler images, comprising:

obtaining, by a processor of a computing system, a raw Doppler image from a series of images of an echocardiogram, received from an echocardiogram machine;

isolating, by the processor, a region of interest from the raw Doppler image, the region of interest including a Doppler image and an electrocardiogram (ECG) signal over at least one heart cycle;

determining, by the processor, a velocity envelope of the Doppler image in the region of interest;

synchronizing, by the processor, the electrocardiogram (ECG) signal with the Doppler image over the at least one heart cycle within the region of interest, wherein the electrocardiogram is extracted by detecting the electrocardiogram (ECG) signal using an energy maximization equation, the energy maximization equation being:

E ( i,j )=λ 1 E continuity ( i,j )+λ 2 E color ( i,j )+λ 3 E gradient ( i,j )+λ 4 E notgray ( i,j )

wherein E(i,j) is an energy value, E continuity (i,j) is a continuity of the electrocardiogram (ECG) signal, E color (i,j) is a color profile of the electrocardiogram (ECG) signal, E gradient (i,j) is a gradient between the color profile of the electrocardiogram (ECG) signal and a background of the image, E notgray (i,j) is RGB value of the electrocardiogram (ECG) signal that is not gray, and each of, λ 1 , λ 2 , λ 3 , and λ 4 is a weighting factor;

calculating, by the processor, a value of a clinical feature from the electrocardiogram (ECG) signal superimposed on the Doppler image; and

comparing, by the processor, the value of the clinical feature with clinical guidelines associated with the clinical feature to determine a diagnosis of a disease.