Automated surface area assessment for dermatologic lesions
A method for assessing a three-dimensional (3D) surface area having one or more lesions is disclosed. The method includes steps of: capturing a two-dimensional (2D) color image and a depth image of the 3D surface area; enhancing contrast of the 2D color image; segmenting the one or more lesions of the 2D color image into one or more segmented lesions; and calculating 3D area of the one or more segmented lesions using information from 2D color image and the depth image.
1. A method for assessing a three-dimensional (3D) surface area having one or more lesions, the method comprising:
capturing a two-dimensional (2D) color image and a depth image of the 3D surface area, the depth image being an image in which each pixel has associated therewith a distance to a corresponding point on the surface being imaged;
enhancing contrast of the 2D color image;
segmenting the one or more lesions of the 2D color image into one or more segmented lesions; and
calculating a 3D surface area of the one or more segmented lesions using enhanced contrast information from the 2D color image and distance information from the depth image.
2. The method of claim 1 further comprising:
calibrating at least one parameter for processing the depth image.
3. The method of claim 1 further comprising:
mapping pixels of the 2D color image to depth pixel of the depth image.
4. The method of claim 1 , the enhancing contrast of the 2D color image further comprising:
applying gamma correction to each of color channels of the 2D color image.
5. The method of claim 1 , the enhancing contrast of the 2D color image further comprising:
applying histogram equalization to the 2D color image.
6. The method of claim 1 , the enhancing contrast of the 2D color image further comprising:
dividing the 2D color image into a plurality of tiled images;
applying the histogram equalization method to each of the tiled images; and
combining the plurality of images into a combined two-image.
7. The method of claim 6 , the enhancing contrast of the 2D color image further comprising:
softening borders of each of the tiled images by applying the bilinear interpolation method.
8. The method of claim 1 , further comprising:
converting a first color space of the 2D color image into a second color space;
adjusting a lightness channel of the 2D color image; and
converting the second color space of the 2D color image to the first color space.
9. The method of claim 1 , the segmenting the surface area further comprising:
segmenting the one or more lesions of the 2D color image into one or more segmented lesions using the watershed transformation.
10. The method of claim 1 , the calculating 3D area of the one or more lesions further comprising:
calculating depth values for each pixel of the depth image; and
calculating the 3D area of one or more lesions considering the depth values.
11. The method of claim 10 , the calculating 3D area of the one or more lesions further comprising:
calculating the 3D area of one or more lesions considering the gradient of the depth values.
12. The method of claim 11 , further comprising:
approximating the gradient of the depth values using a central difference of horizontally aligned pixels and vertically aligned pixels in the depth image.
13. The method of claim 1 , wherein calculating the 3D surface area of the one or more segmented lesions further comprises:
calculating the 3D surface area of the one or more segmented lesions using the following formulas:
Lesion
area
≅
∑
(
i
,
j
)
∈
Lesion
D
(
i
,
j
)
2
D
x
(
i
,
j
)
2
+
D
y
(
i
,
j
)
2
+
1
θ
x
θ
y
D
x
(
i
,
j
)
=
D
(
i
+
1
,
j
)
-
D
(
i
-
1
,
j
)
2
D
(
i
,
j
)
θ
x
D
y
(
i
,
j
)
=
D
(
i
,
j
+
1
)
-
D
(
i
,
j
-
1
)
2
D
(
i
,
j
)
θ
y
where (i, j) are pixel coordinates, D(i, j) is the depth of pixel (i, j), and the camera has a field of view Fx×Fy degrees with a resolution of Rx×Ry pixels and each image pixel corresponds to a field of view of θx×θy degrees where θx=Fx/Rx and θy=Fy/Ry.
14. The method of claim 1 further comprising:
measuring distance to each of pixels of the depth image by measuring the time of flight between the each of pixels and a camera capturing the depth image.
15. The method of claim 1 , further comprising:
converting an RGB color space of the 2D color image into a Lab color space;
adjusting a lightness channel of the 2D color image; and
converting the Lab color space of the 2D color image to the RGB color space.
16. An apparatus for assessing a three-dimensional (3D) surface area having one or more lesions, the apparatus comprising:
a camera capturing a two-dimensional (2D) color image and a depth image of the 3D surface area, the depth image being an image in which each pixel has associated therewith a distance to a corresponding point on the surface being imaged; and
a processor and a memory, wherein the processor is configured to:
enhance contrast of the 2D color image;
segment the one or more lesions of the 2D color image into one or more segmented lesions; and
calculate a 3D surface area of the one or more segmented lesions using enhanced contrast information from the 2D color image and distance information from the depth image.
17. The apparatus of claim 16 , wherein the camera measures distance to each of pixels of the depth image by measuring the time of flight between the each of pixels and the camera.
18. The apparatus of claim 16 , further comprising a user interface with which a user enhances the contrast of the 2D color image.
19. The apparatus of claim 18 , wherein the user selects a type of histogram equalization to enhance the contrast of the 2D color image via the user interface.