Crowdshaping realistic 3D avatars with words
A method for generating a body shape, comprising: receiving one or more linguistic descriptors related to the body shape; retrieving an association between the one or more linguistic descriptors and a body shape; and generating the body shape, based on the association.
1. Computer-implemented method for generating a body shape, the method comprising:
receiving one or more linguistic descriptors related to the body shape;
mapping the one or more linguistic descriptors to coefficients describing the body shape, using a mapping between a linguistic body space and a geometric body space, wherein the linguistic body space is represented in terms of body descriptor words and wherein the geometric body space is represented in terms of coefficient vectors;
generating the body shape, based on the coefficients; and
outputting the generated body shape,
wherein the mapping was learned based on empirical user ratings.
2. The method of claim 1 , wherein the user ratings indicate a degree to which a given linguistic descriptor applies to a given body shape.
3. The method of claim 2 , wherein the degree is selected from a given Likert scale.
4. The method of claim 3 , wherein the mapping is linear.
5. The method of claim 4 , wherein the mapping has been learned using principal component analyses (PCA).
6. The method of claim 1 , used for visualizing word meaning.
7. The method of claim 1 , wherein a slider may be used to set linguistic body descriptors.
8. The method of claim 2 , wherein the body shapes are presented as photos of persons.
9. The method of claim 1 , wherein the body shape is retrieved from a database.
10. The method of claim 1 , wherein a set of similar body shapes is generated.
11. The method of claim 1 , wherein the linguistic descriptors include words not related to body shape.
12. A device for generating a body shape, the device comprising:
a receiving unit for receiving one or more linguistic descriptors related to the body shape;
a mapping unit for mapping the one or more linguistic descriptors to coefficients describing a body shape, using a mapping between a linguistic body space and a geometric body space, wherein the linguistic body space is represented in terms of body descriptor words and wherein the geometric body space is represented in terms of coefficient vectors;
a generating unit for generating the body shape, based on the coefficients; and
an output unit for outputting the generated body shape,
wherein the mapping is learned based on empirical user ratings.
13. The device of claim 12 , wherein the user ratings indicate a degree to which a given linguistic descriptor applies to a given body shape.
14. The device of claim 13 , wherein the degree is selected from a given Likert scale.
15. A computer-implemented method for searching given body shapes in a database, wherein the body shapes are represented in terms of body shape coefficients, the method comprising:
receiving one or more linguistic descriptors related to a body shape;
predicting a vector of word ratings, based on the body shape coefficients of a body, by using an inverted mapping between a linguistic body space and a geometric body space, wherein the linguistic body space is represented in terms of body descriptor words and wherein the geometric body space is represented in terms of coefficient vectors;
matching the one or more linguistic descriptors and the vector of word ratings; and
outputting the body shape, if a match was found.
16. The method according to claim 15 , wherein the mapping was learned based on empirical user ratings.