Hierarchical differential image filters for skin analysis

Patent metadata

Field	Value
Publication number	US-10755089-B2
Application number	US-201916444921-A
Country	US
Kind code	B2
Filing date	Jun 18, 2019
Priority date	May 8, 2016
Publication date	Aug 25, 2020
Grant date	Aug 25, 2020

Abstract

There is provided a framework including systems and methods for analyzing skin parameters from images or videos showing skin. Using a series of Hierarchical Differential Image Filters (HDIF), it becomes possible to detect different skin features such as wrinkles, spots, and roughness. The hierarchical differential image filter computes two enhancements to an image showing skin at two different levels of enhancement, determines a differential image using the two enhancements and computes the skin analysis rating using the differential image. These skin ratings are comparably accurate to actual ratings by dermatologists.

First claim

The invention claimed is:

1. A skin analysis unit comprising at least one processor configured to:
analyze a plurality of related images from a video showing skin using a hierarchical differential image filter to determine a plurality of candidate skin analysis ratings from each image of the plurality of related images, wherein the hierarchical differential image filter computes two enhancements to each image at two different levels of enhancement, determines a differential image using the two enhancements and computes one of the candidate skin analysis ratings from the differential image; and
determine and provide a final skin analysis rating from the plurality of candidate skin analysis ratings.

2. The skin analysis unit of claim 1 wherein the final skin analysis rating is determined by selecting a maximum rating from the plurality of candidate skin analysis ratings.

3. The skin analysis unit of claim 1 wherein the plurality of related images are selected from successive images of a same area of skin from the video.

4. The skin analysis unit of claim 1 wherein the two different levels of enhancement define a skin analysis level configured to determine each of the candidate skin analysis ratings for a specific skin issue.

5. The skin analysis unit of claim 4 configured to perform skin analysis using different skin analysis levels, applying respective hierarchical differential image filters wherein the different skin analysis levels comprise two or more of:
Level 1—facial texture comprising roughness and imperfections;
Level 2—dark spots and small wrinkles; and
Level 3—deep wrinkles and folds;

to determine respective skin analysis ratings for each specific skin issue.

6. The skin analysis unit of claim 1 wherein the hierarchical differential image filter computes the two enhancements by applying a box blur function to each image of the plurality of related images at the two different levels of enhancement.

7. The skin analysis unit of claim 1 wherein each image of the plurality of images is respectively defined by I(x,y) and the hierarchical differential image filter applies an image enhancement function Ω[I(x,y),u] with enhancement level u corresponding to the intensity of the enhancement, with Ω[I(x,y),0]=I(x,y).

8. The skin analysis unit of claim 7 wherein each respective differential image is generally defined by D u,v (x,y), with v>u such that:
D u,v ( x,y )=max(Ω[ I ( x,y ), v )]−Ω[ I ( x,y ), u ],0)  (Eq. I)

wherein a max operation is performed for skin problem areas that are darker than actual skin tone and a −min (negative minimum) operation is performed for skin problem areas that are lighter than actual skin tone.

9. The skin analysis unit of claim 8 configured to analyse each image of the plurality of related images at different skin analysis levels i and i+1 using respective hierarchical differential image filters and wherein the differential image for level i is determined from an adjusted image where skin problem areas of the higher level i+1 are removed from the current level i.

10. The skin analysis unit of claim 1 wherein:
the image enhancement function is a box blurring function having a box width determining a level of enhancement (i) applied to each image of the plurality of related images; and
for each level i, the hierarchical differential image filter applies box widths u i and v i where each respective box width is defined as a respective ratio of a face width of a face in each image of the plurality of related images.

11. The skin analysis unit of claim 10 configured to perform skin analysis using at least one of a plurality of different skin analysis levels, applying respective hierarchical differential image filters wherein the different skin analysis levels are selected from:
Level 1→{u 0 ,v 0 }={0%, 2%};
Level 2→{u 1 ,v 1 }={0%, 5%}; and
Level 3→{u 2 ,v 2 }={7%, 12%}.

12. The skin analysis unit of claim 1 configured to normalize the final skin analysis rating.

13. The skin analysis unit of claim 1 comprising either an on-board camera or a remote camera coupled to the unit and wherein the unit is configured to capture the video using the on-board camera or remote camera.

14. A method of skin analysis comprising:
analyzing a plurality of related images from a video showing skin using a hierarchical differential image filter to determine a plurality of candidate skin analysis ratings from each image of the plurality of related images, wherein the hierarchical differential image filter computes two enhancements to each image at two different levels of enhancement, determines a differential image using the two enhancements and computes one of the candidate skin analysis ratings from the differential image; and
determine and provide a final skin analysis rating from the plurality of candidate skin analysis ratings.

15. The method of claim 14 wherein determining the final skin analysis rating comprises selecting a maximum rating from the plurality of candidate skin analysis ratings.

16. The method of claim 14 comprising selecting the plurality of related images from successive images of a same area of skin from the video.

17. The method of claim 14 wherein the two different levels of enhancement define a skin analysis level configured to determine each of the candidate skin analysis ratings for a specific skin issue.

18. The method of claim 17 comprising performing skin analysis using different skin analysis levels, applying respective hierarchical differential image filters wherein the different skin analysis levels comprise two or more of:
Level 1—facial texture comprising roughness and imperfections;
Level 2—dark spots and small wrinkles; and
Level 3—deep wrinkles and folds;
to determine respective skin analysis ratings for each specific skin issue.

19. The method of claim 14 comprising using the hierarchical differential image filter to compute the two enhancements by applying a box blur function to each image of the plurality of images at the two different levels of enhancement.

20. The method of claim 14 wherein each of the plurality of images of the skin is respectively defined by I(x,y) and wherein using the hierarchical differential image filter applies an image enhancement function Ω[I(x,y),u] with enhancement level u corresponding to the intensity of the enhancement, with Ω[I(x,y),0]=I(x,y).

21. The method of claim 20 wherein each respective differential image is generally defined by D u,v (x,y), with v>u such that:
D u,v ( x,y )=max(Ω[ I ( x,y ), v )]−Ω[ I ( x,y ), u ],0)  (Eq. I)

wherein a max operation is performed for skin problem areas that are darker than actual skin tone and a −min (negative minimum) operation is performed for skin problem areas that are lighter than actual skin tone.

22. The method of claim 21 comprising analysing each image of the plurality of related images at different skin analysis levels i and i+1 using respective hierarchical differential image filters and wherein the differential image for level i is determined from an adjusted image where skin problem areas of the higher level i+1 are removed from the current level i.

23. The method of claim 14 wherein:
the image enhancement function is a box blurring function having a box width determining a level of enhancement (i) applied to each image of the plurality of related images; and
for each level i, the hierarchical differential image filter applies box widths u i

Assignees

• L'Oréal
• Oreal

Inventors

• Aarabi Parham
• Zhang Angel Jing Yi

Classifications

• G06V10/52

  Scale-space analysis, e.g. wavelet analysis (multi-scale boundary representations G06V10/42) · CPC title

• G06V10/443

  by matching or filtering · CPC title

• G06V40/10 Primary

  Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands · CPC title

• G06V30/2504

  Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches · CPC title

• G06V2201/03

  Recognition of patterns in medical or anatomical images · CPC title