Method and system for visualizing dyed hair image
Disclosed are a method and system for visualizing a dyed hair image. The method of visualizing dyeing hair includes separating a hair area from an image related to a dyeing target, detecting a color characteristic of the hair area, determining a target color corresponding to the color characteristic by using a dataset comprising color difference values of dyeing hair using a hairdye product, performing a color transfer using the target color on the hair area, and merging the hair area on which the color transfer using the target color as been performed with the image.
1 . A method of visualizing dyeing hair, which is performed in a computer device,
wherein the computer device comprises at least one processor configured to execute computer-readable instructions included in memory, and
the method of visualizing dyeing hair comprises:
separating, by the at least one processor, a hair area from an image related to a dyeing target;
detecting, by the at least one processor, a color characteristic of the hair area;
determining, by the at least one processor, a target color corresponding to the color characteristic by using a dataset comprising color difference values of dyeing hair using a hairdye product;
performing, by the at least one processor, a color transfer using the target color on the hair area; and
merging, by the at least one processor, the hair area on which the color transfer using the target color as been performed with the image,
wherein:
detecting the color characteristic comprises extracting a brightness level and undertone level of a hair color based on a color value of pixels that constitute the hair area, and
performing the color transfer comprises performing the color transfer for transferring a representative color to the target color by considering the brightness level and the undertone level as a representative color, wherein a distribution of pixels brightness deviations is modified based on a brightness of the target color.
2 . The method of claim 1 , wherein separating the hair area comprises:
applying white balancing to the image; and
cutting, as an image based on a face area, the image to which the white balancing has been applied.
3 . The method of claim 1 , wherein separating the hair area comprises separating the hair area from the image along with a face skin area through image segmentation.
4 . The method of claim 1 , wherein detecting the color characteristic comprises calculating a brightness level and undertone level of a hair color based on a distribution of lab color values of pixels that constitute the hair area.
5 . The method of claim 1 , wherein:
the dataset comprises color values of reference hair measured by a spectophotometer after the reference hair, categorized by each of initial hair brightness levels, was dyed using the hair dye product, and
determining the target color comprises determining the target color that corresponds to the color characteristic and that has been predicted based on the results of the dyeing by using the dataset through linear interpolation.
6 . The method of claim 1 , wherein:
the dataset comprises predefined reference color values of reference hair corresponding to dying results after the reference hair, categorized by each of initial hair brightness levels, was dyed using the hair dye products, and
determining the target color comprises determining, as the target color, a reference color value among the predefined reference color values most similar to the color characteristic by using the dataset.
7 . The method of claim 1 , wherein merging the hair area comprises merging, with the image, an edge area of the hair area on which the color transfer using the target color has been performed, after assigning transparency to the edge area.
8 . The method of claim 1 , further comprising, by the at least one processor, extracting a skin color based on a color value of pixels that constitute a face skin area separated from the image,
wherein merging the hair area comprises merging the hair area on which the color transfer using the target color has been performed, with an area in which a color difference with the skin color is at least a predetermined level in the image.
9 . A computer program which is stored in a computer-readable recording medium in order to execute a method of visualizing dyeing hair in a computer,
wherein the method of visualizing dyeing hair comprises:
separating a hair area from an image related to a dyeing target;
detecting a color characteristic of the hair area;
determining a target color corresponding to the color characteristic by using a dataset comprising color difference values of dyeing hair using a hairdye product;
performing a color transfer using the target color on the hair area; and
merging the hair area on which the color transfer using the target color as been performed with the image,
wherein,
the dataset comprises predefined reference color values of reference hair corresponding to dying results after the reference hair, categorized by each of initial hair brightness levels, was dyed using the hair dye product, and
determining the target color comprises determining, as the target color, a reference color value among the predefined reference color values most similar to the color characteristic by using the dataset.
10 . A system for visualizing dyeing hair, which is implemented as a computer, the system comprising:
at least one processor configured to execute computer-readable instructions included in memory,
the at least one processor is configured to:
separate a hair area from an image related to a dyeing target,
detect a color characteristic of the hair area,
determine a target color corresponding to the color characteristic by using a dataset comprising color difference values of dyeing hair using a hairdye product,
perform a color transfer using the target color on the hair area, and
merge the hair area on which the color transfer using the target color as been performed with the image,
wherein,
the dataset comprises color values of reference hair measured by a spectrophotometer after the reference hair, categorized by each of initial hair brightness levels, was dyed using the hair dye product, and
the at least one processor is configured to determine the target color that corresponds to the color characteristic and that has been predicted based on the results of the dyeing by using the dataset through linear interpolation.
11 . The system of claim 10 , wherein the at least one processor is configured to calculate a brightness level and undertone level of a hair color based on a distribution of lab color values of pixels that constitute the hair area.
12 . The system of claim 10 , wherein the at least one processor is configured to merge, with the image, an edge area of the hair area on which the color transfer using the target color has been performed, after assigning transparency to the edge area.
13 . The system of claim 10 , wherein the at least one processor is configured to:
extract a skin color based on a color value of pixels that constitute a face skin area separated from the image, and
merge the hair area on which the color transfer using the target color has been performed, with an area in which a color difference with the skin color is at least a predetermined level in the image.