Method and apparatus for enhancing and correcting digital images
AssigneeGE Medical Systems Global Technology Company, LLC
InventorGopal B. Avinash
Life of this patent
Estimating the term…
Abstract
A digital image processing method includes the steps of acquiring (60) pixel data which defines a digital image (36), identifying (66) structural regions (48) within the digital image, resulting in a structure mask showing the structural regions and non-structural regions (50), enhancement filtering (68-76) the digital image using the structure mask, and correcting (78) for intensity non-uniformities in the digital image, resulting in an enhanced, corrected digital image. An apparatus for carrying out the method includes a signal processing circuit configured to receive signals representing the internal features and to perform the steps (60-78) of the method.
Claims9
A method for enhancing and correcting a digital image made up of a pixel array, the method comprising the steps of:
- a) acquiring pixel data which defines a digital image of internal features of a physical subject;
- b) identifying structural regions within the digital image defined by the pixel data, resulting in a structure mask showing the structural regions and non-structural regions;
- c) enhancement filtering the digital image using the structure mask, resulting in a first filtered image;
- and
- d) correcting for intensity non-uniformities in the first filtered image, resulting in an enhanced, corrected digital image, the correcting step including
- d1) reading in image data representing the first filtered image;
- d2) determining a first threshold value, T;
- d3) reducing the pixel array using a shrink parameter, S, resulting in a second array;
- d4) changing values of pixels of the second array based on comparison of the values with the first threshold value, resulting in a third array;
- d5) transforming and filtering the second array and the third array, resulting in a fourth array and a fifth array;
- d6) maximizing the fourth array and the fifth array by changing values of pixels of the fourth array and the fifth array, resulting in a sixth array and a seventh array;
- d7) determining a distortion function based on the sixth array and the seventh array;
- d8) computing a corrected function from the distortion function and the image data;
- and
- d9) rescaling the corrected function back to an original intensity range.
An apparatus for enhancing and correcting a digital image made up of a pixel array, wherein the digital image represents an image of internal features of a physical subject, the apparatus comprising:
- a signal acquisition circuit configured to receive and transmit signals representing the internal features;
- and
- a signal processing circuit configured to receive the signals, acquire pixel data by digitizing the signals, and to enhance and correct the digital image by identifying structural regions within the digital image defined by the pixel data, resulting in a structure mask showing the structural regions and non-structural regions, enhancement filtering the digital image using the structure mask, resulting in a first filtered image, the enhancement filtering including the steps of
- a) orientation smoothing the structural regions of the digital image based on the structure mask, resulting in a second filtered image;
- b) homogenization smoothing the non-structural regions of the second filtered image in order to blend features of the non-structural regions into an environment surrounding the structural regions, resulting in a third filtered image;
- c) orientation sharpening the structural regions of the third filtered image, resulting in a fourth filtered image;
- d) renormalizing the fourth filtered image, resulting in a fifth filtered image;
- and
- e) texture blending the non-structural regions of the fifth filtered image, and
- correcting for intensity non-uniformities in the first filtered image, resulting in an enhanced, corrected digital image.
A method for enhancing and correcting a digital image made up of a pixel array, the method comprising the steps of:
- a) acquiring pixel data which defines a digital image of internal features of a physical subject;
- b) identifying structural regions within the digital image defined by the pixel data, resulting in a structure mask showing the structural regions and non-structural regions;
- c) enhancement filtering the digital image using the structure mask, resulting in a first filtered image, the enhancement filtering including the steps of
- c1) orientation smoothing the structural regions of the digital image based on the structure mask, resulting in a second filtered image;
- c2) homogenization smoothing the non-structural regions of the second filtered image in order to blend features of the non-structural regions into an environment surrounding the structural regions, resulting in a third filtered image;
- c3) orientation sharpening the structural regions of the third filtered image, resulting in a fourth filtered image;
- c4) renormalizing the fourth filtered image, resulting in a fifth filtered image;
- and
- c5) texture blending the non-structural regions of the fifth filtered image, resulting in the first filtered image;
- and
- d) correcting for intensity non-uniformities in the first filtered image, resulting in an enhanced, corrected digital image.
Dependencies are inferred from the claim text.