Apparatus and method of guided neural network model for image processing
The present disclosure provides an apparatus and method of guided neural network model for image processing. An apparatus may comprise a guidance map generator, a synthesis network and an accelerator. The guidance map generator may receive a first image as a content image and a second image as a style image, and generate a first plurality of guidance maps and a second plurality of guidance maps, respectively from the first image and the second image. The synthesis network may synthesize the first plurality of guidance maps and the second plurality of guidance maps to determine guidance information. The accelerator may generate an output image by applying the style of the second image to the first image based on the guidance information.
1. An apparatus comprising:
one or more processors to:
receive a first image as a content image and a second image as a style image;
generate a first plurality of guidance maps and a second plurality of guidance maps, respectively, from the first image and the second image;
synthesize the first plurality of guidance maps and the second plurality of guidance maps to determine guidance information; and
generate an output image by applying the style of the second image to the first image based on the guidance information.
2. The apparatus of claim 1 , wherein the one or more processors are further to:
generate a first neural guidance map and a second neural guidance map, respectively, from the first image and the second image; and/or
generate a first semantic guidance map and a second semantic guidance map, respectively, from the first image and the second image; and/or
generate a first position guidance map and a second position guidance map, respectively, from the first image and the second image.
3. The apparatus of claim 2 , wherein the one or more processors are further to:
generate the first neural guidance map by removing a style from the first image; and/or
generate the second neural guidance map by removing a style from the second image; and/or
determine boundaries of parts of each of the first image and the second image through semantic parsing; and/or
detect landmarks within the boundaries of the parts that represent facial components; and/or
fit boundary curves of the facial components based on the detected landmarks; and/or
generate component masks based on the fitted boundary curves; and/or
generate the first semantic guidance map and the second semantic guidance map by combining the component masks with the boundaries of the parts.
4. The apparatus of claim 2 , wherein the one or more processors are further to:
fit boundary curves of eyeballs; and/or
generate the first semantic guidance map and the second semantic guidance map by combining the component masks with the boundary curves of eyeballs.
5. The apparatus of claim 2 , wherein the one or more processors are further to:
determine boundaries of parts of each of the first image and the second image through semantic parsing; and/or
determine a score for each pixel within a boundary of a part based on a minimal distance of the pixel from the boundary; and/or
generate the first position guidance map and the second position guidance map based on scores of pixels.
6. The apparatus of claim 1 , wherein the one or more processors are further to synthesize the first and second neural guidance maps, the first and second semantic guidance maps, and the first and second position guidance maps to determine the guidance information.
7. The apparatus of claim 1 , wherein the one or more processors are coupled to a memory, the one or more processors comprising one or more graphics processors or one or more application processors.
8. A method comprising:
receiving, by one or more processors, a first image as a content image and a second image as a style image;
generating a first plurality of guidance maps and a second plurality of guidance maps, respectively, from the first image and the second image;
synthesizing the first plurality of guidance maps and the second plurality of guidance maps to determine guidance information; and
generating an output image by applying the style of the second image to the first image based on the guidance information.
9. The method of claim 8 , wherein generating the first plurality of guidance maps and the second plurality of guidance maps comprising:
generating a first neural guidance map and a second neural guidance map, respectively, from the first image and the second image; and/or
generating a first semantic guidance map and a second semantic guidance map, respectively, from the first image and the second image; and/or
generating a first position guidance map and a second position guidance map, respectively, from the first image and the second image.
10. The method of claim 9 , wherein generating the first neural guidance map and the second neural guidance map comprises:
generating the first neural guidance map by removing a style from the first image; and/or
generating the second neural guidance map by removing a style from the second image.
11. The method of claim 9 , wherein generating the first semantic guidance map and the second semantic guidance map comprises:
determining boundaries of parts of each of the first image and the second image through semantic parsing;
detecting landmarks within the boundaries of the parts that represent facial components; and/or
fitting boundary curves of the facial components based on the detected landmarks; and/or
generating component masks based on the fitted boundary curves; and/or
generating the first semantic guidance map and the second semantic guidance map by combining the component masks with the boundaries of the parts; and/or
fitting boundary curves of eyeballs; and/or,
generating the first semantic guidance map and the second semantic guidance map by combining the component masks with the boundary curves of eyeballs.
12. The method of claim 9 , wherein generating the first position guidance map and the second position guidance map comprises:
determining boundaries of parts of each of the first image and the second image through semantic parsing;
determining a score for each pixel within a boundary of a part based on a minimal distance of the pixel from the boundary; and
generating the first position guidance map and the second position guidance map based on scores of pixels,
wherein synthesizing the first plurality of guidance maps and the second plurality of guidance maps comprises synthesizing the first and second neural guidance maps, the first and second semantic guidance maps, and the first and second position guidance maps to determine the guidance information.
13. The method of claim 8 , wherein each of the style of the first image and the style of the second image comprises a portrait style or a lighting style, wherein the parts comprise a face, a body, or hair.
14. The method of claim 8 , wherein the one or more processors are coupled to a memory, the one or more processors comprising one or more graphics processors or one or more application processors.
15. At least one non-transitory computer-readable medium comprising instructions which, when executed, cause a computing device to perform operations comprising:
receiving, by one or more processors, a first image as a content image and a second image as a style image;
generating a first plurality of guidance maps and a second plurality of guidance maps, respectively, from the first image and the second image;
synthesizing the first plurality of guidance maps and the second plurality of guidance maps to determine guidance information; and
generating an output image by applying the style of the second image to the first image based on the guidance information.
16. The computer-readable medium of claim 15 , wherein the operations further comprise:
generating a first neural guidance map and a second neural guidance map, respectively, from the first image and the second image; and/or
generating a first semantic guidance map and a second semantic guidance map, respectively, from the first image and the second image; and/or
generating a first position guidance map and a second position guidance map, respectively, from the first image and the second image.
17. The computer-readable medium of claim 15 , wherein the operations further comprise:
generating the first neural guidance map by removing a style from the first image;
and/or generating the second neural guidance map by removing a style from the second image.
18. The computer-readable medium of claim 15 , wherein the operations further comprise:
determining boundaries of parts of each of the first image and the second image through semantic parsing;
detecting landmarks within the boundaries of the parts that represent facial components; and/or
fitting boundary curves of the facial components based on the detected landmarks; and/or
generating component masks based on the fitted boundary curves; and/or
generating the first semantic guidance map and the second semantic guidance map by combining the component masks with the boundaries of the parts; and/or
fitting boundary curves of eyeballs; and/or;
generating the first semantic guidance map and the second semantic guidance map by combining the component masks with the boundary curves of eyeballs.
19. The computer-readable medium of claim 15 , wherein the operations further comprise:
determining boundaries of parts of each of the first image and the second image through semantic parsing;
determining a score for each pixel within a boundary of a part based on a minimal distance of the pixel from the boundary; and
generating the first position guidance map and the second position guidance map based on scores of pixels,
wherein synthesizing the first plurality of guidance maps and the second plurality of guidance maps comprises synthesizing the first and second neural guidance maps, the first and second semantic guidance maps, and the first and second position guidance maps to determine the guidance information.
20. The computer-readable medium of claim 15 , wherein each of the style of the first image and the style of the second image comprises a portrait style or a lighting style, wherein the parts comprise a face, a body, or hair.