Cartoon Face Generation

• US 20090087035A1
• Filed: 10/02/2007
• Published: 04/02/2009
• Est. Priority Date: 10/02/2007
• Status: Active Grant

• Current Assignee: Zhigu Holdings Limited
• Original Assignee: Microsoft Corporation
• Inventors: Zou, Xin, Wen, Fang, Xu, Ying-Qing, Liang, Lin
• Granted Patent: US 8,437,514 B2
• Time in Patent Office: Days
• US Class Current: 382/118
• CPC Class Codes: G06T 11/00 2D [Two Dimensional] image generation
• CPC Class Codes: G06V 40/168 Feature extraction; Face representation

Abstract

A face cartooning system is described. In one implementation, the system generates an attractive cartoon face or graphic of a user'"'"'s facial image. The system extracts facial features separately and applies pixel-based techniques customized to each facial feature. The style of cartoon face achieved resembles the likeness of the user more than cartoons generated by conventional vector-based cartooning techniques. The cartoon faces thus achieved provide an attractive facial appearance and thus have wide applicability in art, gaming, and messaging applications in which a pleasing degree of realism is desirable without exaggerated comedy or caricature.

Claims

1. A method, comprising:

extracting visual features of a face portrayed in an image;

assigning pixels of each visual feature into clusters according to an attribute;

shifting pixel values across each visual feature based on one of the clusters in each visual feature; and

recombining the visual features to form a cartoon of the face.

2. The method as recited in claim 1, wherein extracting the visual features includes:

detecting a face;

aligning the face to map some of the visual features with points; and

extracting the visual features and associated alpha-matte masks.

3. The method as recited in claim 1, wherein the features include one of facial skin, facial shadow, eyes, brows, nose, cheeks, lips, and inner mouth.

4. The method as recited in claim 2, wherein mapping the visual features includes applying a Bayesian Tangent Shape Model (BSTM).

5. The method as recited in claim 1, wherein assigning pixels of each visual feature into clusters further includes:

grouping the pixels into the clusters according a lightness attribute; and

selecting one of the clusters to represent the color of the visual feature.

6. The method as recited in claim 5, further comprising selecting the second lightest cluster to represent the color of the visual feature, when the visual feature comprises a skin region.

7. The method as recited in claim 5, further comprising:

propagating the color of the selected cluster across the pixels of the visual feature; and

shifting the lightness and chroma of the pixels of the visual feature to achieve a cartoon style.

8. The method as recited in claim 1, further comprising learning an affine transformation from a cheek shape of a reference face to a cheek shape of the face in the image to produce a forehead and ear geometry for the face in the image based on a forehead and ear geometry of the reference face.

9. The method as recited in claim 1, wherein extracting visual features includes:

extracting a shadow region of the face;

clustering pixels in the shadow region into groups according to a lightness attribute of the pixels; and

replacing a color of each pixel in each group with a mean color of the group.

10. The method as recited in claim 1, further comprising adopting a different cartooning technique for each different visual feature of the face.

11. The method as recited in claim 1, further comprising adopting a same cartooning technique for each different visual feature of the face, and using different parameters for the cartooning technique for each visual feature.

12. The method as recited in claim 1, further comprising:

applying a pixel-based cartooning technique to an extracted skin region of the face;

applying a pixel-based cartooning technique to an extracted head shape feature of the face; and

combining the skin region of the face and the head shape feature of the face to create a base for recombining the remaining visual features of the face.

13. The method as recited in claim 12, further comprising combining the base with a shadow visual feature of the face to form a second base for recombining the subsequent remaining visual features of the face.

14. The method as recited in claim 1, further comprising combining accessories with the cartoon of the face, wherein the accessories include one of a neck, hair, hat, eyeglasses, facial hair, and clothing.

15. A system, comprising:

a computing device;

a cartooning engine in the computing device;

a decomposition engine in the cartooning engine to extract visual features from a face portrayed in an image;

a cartoonizer to assign pixels of each visual feature into clusters according to an attribute and shift pixel values across each visual feature based on one of the clusters in each visual feature; and

a compositor to recombine the visual features using an alpha-matte technique to form a cartoon of the face.

16. The system as recited in claim 15, further comprising a face processor, including a face detector, pre-processing components to enhance contrast and normalize color, and an alignment engine to align the face in the image for mapping facial features.

17. The system as recited in claim 16, further comprising a user interface for a user to interactively assist mapping the facial features.

18. The system as recited in claim 15, wherein the cartoonizer renders each of the visual features separately into a cartoon style;

and

wherein the visual features include a skin feature, a shadow feature, a facial geometry feature, an eyebrows features, an eyes feature, a lips feature, and an inner-mouth feature.

19. The system as recited in claim 15, further comprising an accessories engine, for adding user-selected pre-drawn visual features to the cartoon of the face.

20. A system, comprising:

means for extracting visual features from a facial image; and

means for separately rendering each visual feature into a cartoon style based on pixel clustering.