Image enhancement method
View Patent ↗An image enhancement method. In the first stage, an image described in first image format is received. Then, each pixel is assessed with respect to its location on the edge of the image, and the value of the pixel in second image format is assessed with respect to its falling within the range of skin tones. A first smooth filter mask is dynamically generated if the pixel is not on the edge and the value does not fall into the range of skin tones, and a smooth process is performed on the pixel according to the first smooth filter mask. In the second stage, a second smooth filter mask is dynamically generated for each pixel in the image, and a general smooth process is performed on the pixel according to the second smooth filter mask.
1. An image enhancement method, comprising the steps of:
receiving an image described in first image format; and, for each pixel in the image:
determining whether the pixel is on the edge;
determining whether the value of the pixel in a second image format falls into the range of skin tones;
generating a first smooth filter mask if the pixel is not on the edge and the value does not fall into the range of skin tones; and
performing a smooth process on the pixel according to the first smooth filter mask.
2. The method as claimed in claim 1 further calculating the value of the pixel in second image format.
3. The method as claimed in claim 1 further comprising the steps of, for each pixel in the image:
generating a second smooth filter mask; and
performing a general smooth process on the pixel according to the second smooth filter mask.
4. The method as claimed in claim 3 wherein all positions in the second smooth filter mask are set as valid.
5. The method as claimed in claim 3 wherein the method of generating the second smooth filter mask comprises the steps of:
generating an empty second smooth filter mask; and
setting the weight value of the position corresponding to each of the adjacent pixels in the second smooth filter mask according to the difference of the value with second image format between the adjacent pixel and the pixel and the distance between the adjacent pixel and the pixel.
6. The method as claimed in claim 5 wherein the method of performing the general smooth process on the pixel according to the second smooth filter mask comprises the steps of:
performing a weighted process on all values with first image format of the adjacent pixels and the value with first image format of the pixel according to the weight values set in the second smooth filter mask, thus a new weighted value is acquired; and
replacing the value with first image format of the pixel by the new weighted value.
7. The method as claimed in claim 1 wherein the first image format is RGB format.
8. The method as claimed in claim 1 wherein the second image format is HSV format.
9. The method as claimed in claim 1 wherein an edge detection method is employed to determine whether the pixel is on the edge.
10. The method as claimed in claim 1 wherein the method of generating the first smooth filter mask comprises the steps of:
generating an empty first smooth filter mask;
determining whether the value of each of the adjacent pixels adjacent to the pixel in second image format falls into the range of skin tones; and
setting the position corresponding to the adjacent pixel in the first smooth filter mask as valid.
11. The method as claimed in claim 10 wherein the method of performing the smooth process on the pixel according to the first smooth filter mask comprises the steps of:
giving a fixed weight value to the positions recorded as valid in the first smooth filter mask;
calculating the weight value corresponding to the pixel according to the fixed weight values corresponding to the positions set as valid in the first smooth filter mask;
performing a weighted process on the values with first image format of the adjacent pixels corresponding to the positions set as valid in the first smooth filter mask and the value with first image format of the pixel according to the fixed weight values corresponding to the adjacent pixels and the weight value corresponding to the pixel, thus acquiring a weighted value; and
replacing the value with first image format of the pixel by the weighted value.
12. The method as claimed in claim 1 wherein the method of generating the first smooth filter mask comprises the steps of:
generating an empty first smooth filter mask; and
setting the weight value of the position corresponding to each of the adjacent pixels in the first smooth filter mask according to the difference of the value with second image format between the adjacent pixel and a target skin tones and the distance between the adjacent pixel and the pixel.
13. The method as claimed in claim 12 wherein the method of performing the smooth process on the pixel according to the first smooth filter mask comprises the steps of:
performing a weighted process on all values with first image format of the adjacent pixels and the value with first image format of the pixel according to the weight values set in the first smooth filter mask, thus a weighted value is acquired; and
replacing the value with first image format of the pixel by the weighted value.