Method for automatic facial impression transformation, recording medium and device for performing the method
A method for automatic facial impression transformation includes extracting landmark points for elements of a target face whose facial impression is to be transformed as well as distance vectors respectively representing distances of the landmark points, comparing the distance vectors to select a learning data set similar to the target face from a database, extracting landmark points and distance vectors from the learning data set, transforming a local feature of the target face based on the landmark points of the learning data set and score data for a facial impression, and transforming a global feature of the target face based on the distance vectors of the learning data set and the score data for the facial impression. Accordingly, a facial impression may be transformed in various ways while keeping an identity of a corresponding person.
1. A method for automatic facial impression transformation, comprising:
extracting landmark points for elements of a target face whose facial impression is to be transformed as well as distance vectors respectively representing distances of the landmark points;
acquiring, from a database, a learning data set associated with a facial shape similar to the target face based on the distance vectors;
extracting landmark points and distance vectors from the learning data set;
determining a first distance vector of the target face used for transforming a local feature of the target face based on distance vectors between landmark points determining sizes and shapes of face elements of the learning data set, and determining a second distance vector of the target face used for transforming a global feature of the target face based on distance vectors between landmark points determining arrangement of the face elements of the learning data set; and
transforming the facial impression of the target face based on at least one of the first distance vector and the second distance vector,
wherein the learning data set used is the same for the transforming of the local feature of the target face and the transforming of the global feature of the target face.
2. The method of claim 1 , wherein the transforming of the facial impression of the target face comprises:
generating a target function based on the landmark points of the learning data set and score data of a facial impression; and
determining shapes and sizes of the elements of the target face by moving landmark points associated with the first distance vector, based on the target function.
3. The method of claim 1 , wherein the, transforming of the facial impression of the target face comprises:
generating a target function based on the distance vectors of the learning data set and score data of a facial impression; and
determining locations and angles of the elements of the target face by moving landmark points associated with the second distance vector, based on the target function.
4. The method of claim 1 , further comprising receiving a user input of selecting a target facial impression type among a plurality of facial impression types,
wherein the acquiring of the learning data set from the database comprises acquiring a learning data set associated with the target facial impression type.
5. The method of claim 1 , further comprising continuously outputting frames of the moving picture including the target face whose facial impression is transformed.
6. The method of claim 1 , further comprising acquiring a target face whose facial impression is to be transformed from a frame of a moving picture.
7. The method of claim 6 , wherein the target face is an expressionless photograph staring at front.
8. The method of claim 1 , further comprising forming a database storing learning data.
9. The method of claim 8 , wherein the forming of the database storing the learning data comprises:
acquiring expressionless front facial images; and
extracting landmark points and distance vectors from the expressionless front facial images.
10. The method of claim 8 , wherein the forming of the database storing the learning data comprises collecting and accumulating score data for facial impressions.
11. A non-transitory computer-readable recording medium, on which a computer program for executing the method for automatic facial impression transformation according to claim 1 is recorded.
12. A device for automatic facial impression transformation, comprising:
a facial feature extracting unit configured to extract landmark points for elements of a target face whose facial impression is to be transformed, included in a frame of a moving picture, as well as distance vectors respectively representing distances of the landmark points; and
a facial impression transforming unit configured to acquire, from a database, a learning data set associated with a facial shape most similar to the target face based on the distance vectors, extract landmark points and distance vectors from the learning data set, and transform the facial impression of the target face based on the extracted landmark points and distance vectors of the learning data set,
wherein the facial impression transforming unit comprises:
a local feature transforming unit configured to determine a first distance vector of the target face used for transforming a local feature of the target face, based on distance vectors between landmark points determining sizes and shapes of face elements of the learning data set; and
a global feature transforming unit configured to determine a second distance vector of the target face used for transforming a global feature of the target face, based on distance vectors between landmark points determining arrangement of face elements of the learning data set, and
the facial impression of the target face is transformed based on at least one of the first distance vector and the second distance vector,
wherein the learning data set used is the same for the transforming of the local feature of the target face and the transforming of the global feature of the target face.