System and method for image segmentation and digital analysis for clinical trial scoring in skin disease

Patent metadata

Field	Value
Publication number	US-11244456-B2
Application number	US-201816753582-A
Country	US
Kind code	B2
Filing date	Oct 3, 2018
Priority date	Oct 3, 2017
Publication date	Feb 8, 2022
Grant date	Feb 8, 2022

Abstract

Disclosed are systems and methods for clinical trial assessment of skin disease treatment. The disclosure includes obtaining a series of digital images over a period of time, wherein each digital image includes an affected area of the subject; identifying characteristic morphologies and lesions in the affected area of the subject in each of the digital images; classifying each of the detected and segmented morphologies and lesions into one or more identified categories for each of the digital images; assigning a global score to each of the digital images based on a count of the detected and segmented characteristic morphologies and lesions in each of the one or more identified categories; analyzing the global scores of each of the digital images; and making an assessment of the clinical trial based on the analysis of the global scores of each of the digital images.

First claim

What is claimed is:

1. A computer-implemented method for assessment of a skin disease comprising:
obtaining, by a computer, a series of digital images over a period of time of at least a portion of a subject, wherein each of the digital images includes an affected area of the subject;
identifying, by the computer, characteristic morphologies and lesions in the affected area of the subject in each of the digital images, wherein said identification comprises detection and segmentation of the characteristic morphologies and lesions in each of the digital images;
classifying, by the computer, each of the detected and segmented characteristic morphologies and lesions into one or more identified categories for each of the digital images;
assigning, by the computer, a global score to each of the digital images based on a count of the detected and segmented characteristic morphologies and lesions in each of the one or more identified categories;
analyzing, by the computer, the global scores of each of the digital images; and
making, by the computer, an assessment based on the analysis of the global scores of each of the digital images,
wherein segmentation of the characteristic morphologies and lesions in each of the digital images comprises:
performing color space normalization on each of the series of digital images to determine one or more hue components for each digital image of the series of digital images;
performing morphological opening and closing on the one or more hue components to enhance structural features for each digital image of the series of digital images and to create a morphological opening image and a morphological closing image for each digital image of the series of digital images;
performing iterative thresholding on each of the morphological opening image and the morphological closing image for each digital image of the series of digital images; and
producing a composite image for each digital image of the series of digital images by taking an average of intensity values of corresponding pixels of the morphological opening image and the morphological closing image for each digital image of the series of digital images.

2. The method of claim 1 , further comprising performing color space normalization on each of the series of digital images.

3. The method of claim 2 , wherein performing color space normalization on each of the series of digital images comprises converting each of the digital images into its hue-saturation-value (HSV) color format representation.

4. The method of claim 1 , wherein performing color space normalization on each of the series of digital images comprises converting each of the digital images into its hue-saturation-value (HSV) color format representation.

5. The method of claim 1 , wherein classifying, by the computer, each of the detected and segmented characteristic morphologies and lesions into one or more identified categories for each of the digital images comprises determining a plurality of textural based features of each composite image and classifying each of the detected and segmented characteristic morphologies and lesions into the one or more identified categories based on the determined textural based features.

6. The method of claim 5 , wherein the determined textural based features comprise one or more discrete wavelet frames (DWF) and one or more gray-level co-occurrence matrix descriptors (GLCM).

7. The method of claim 6 , wherein the determined textural based features comprise nine DWF features and eight GLCM features.

8. A computer-implemented method for assessment of temporal changes of a skin disease of a subject comprising:
obtaining a first digital image at a first time, wherein at least a portion of the first digital image comprises an affected area of the subject;
identifying, by a computer, characteristic morphologies and lesions in the affected area of the subject in the first digital image, wherein said identification comprises detection and segmentation of the characteristic morphologies and lesions in the first digital image;
classifying, by the computer, each of the detected and segmented characteristic morphologies and lesions into one or more identified categories for the first digital image;
assigning, by the computer, a global score to the first digital image based on a count of the detected and segmented characteristic morphologies and lesions in each of the one or more identified categories;
obtaining a second digital image of the at least the portion of the affected area, wherein the second digital image is captured at a second time that is after the first digital image was captured;
identifying, by a computer, characteristic morphologies and lesions in the affected area of the subject in the second digital image, wherein said identification comprises detection and segmentation of the characteristic morphologies and lesions in the second digital image;
classifying, by the computer, each of the detected and segmented characteristic morphologies and lesions into one or more identified categories for the second digital image;
assigning, by the computer, a global score to the second digital image based on a count of the detected and segmented characteristic morphologies and lesions in each of the one or more identified categories;
analyzing, by the computer, the global scores of each of the first and second digital images; and
making, by the computer, an assessment based on the analysis of the global scores of each of the first and second digital images, wherein segmentation of the characteristic morphologies and lesions in each of the first and second digital images comprises:
performing color space normalization on each of the first and second digital images to determine one or more hue components for each of the first and second digital images;
performing morphological opening and closing on the one or more hue components to enhance structural features for each of the first and second digital images and to create a morphological opening image and a morphological closing image for each of the first and second digital images;
performing iterative thresholding on each of the morphological opening image and the morphological closing image for each of the first and second digital images; and
producing a composite image for each of the first and second digital images by taking an average of intensity values of corresponding pixels of the morphological opening image and the morphological closing image for each of the first and second digital images.

9. The method of claim 8 , further comprising performing color space normalization on each of the first and second digital images.

10. The method of claim 9 , wherein performing color space normalization on each of the first and second digital images comprises converting each of the digital images into its hue-saturation-value (HSV) color format representation.

11. The method of claim 8 , wherein performing color space normalization on each of the first and second digital images comprises converting each of the first and second digital images into its hue-saturation-value (HSV) color format representation.

12. The method of claim 8 , wherein classifying, by the computer, each of the detected and segmented characteristic morphologies and lesions into one or more identified categories for each of the first and second digital images comprises determining a plurality of textural based features of each composite image and classifying each of the detected and segmented characteristic morphologies and lesions into the one or more identified categories based on the determined textural based features.

13. The method of claim

Assignees

• Ohio State Innovation Foundation

Inventors

• Kaffenberger Benjamin
• Gurcan Metin Nafi

Classifications

• A61B5/444 Primary

  Evaluating skin marks, e.g. mole, nevi, tumour, scar · CPC title

• G06T7/0016 Primary

  involving temporal comparison · CPC title

• G06F18/2431

  Multiple classes · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title