System and method for multimodal emotion recognition
Systems, methods, apparatuses, and computer program products for providing multimodal emotion recognition. The method may include receiving raw input from an input source. The method may also include extracting one or more feature vectors from the raw input. The method may further include determining an effectiveness of the one or more feature vectors. Further, the method may include performing, based on the determination, multiplicative fusion processing on the one or more feature vectors. The method may also include predicting, based on results of the multiplicative fusion processing, one or more emotions of the input source.
1. A method, comprising:
receiving raw input from an input source;
extracting one or more feature vectors from the raw input;
determining an effectiveness of the one or more feature vectors based at least in part on
computing a correlation score for the one or more feature vectors,
checking the computed correlation score against a predetermined threshold, and
a signal noise level,
wherein when the correlation score is above the threshold, the one or more feature vectors is ineffective, and
wherein when the correlation score is below the threshold, the one or more feature vectors is effective;
performing, based on the determination, multiplicative fusion processing on the one or more feature vectors; and
predicting, based on results of the multiplicative fusion processing, one or more emotions of the input source.
2. The method according to claim 1 , wherein the raw input comprises one or more modalities.
3. The method according to claim 1 , wherein the multiplicative fusion processing comprises:
combining the one or more feature vectors with another one or more feature vectors;
boosting one or more of the one or more feature vectors; and
suppressing one or more of the one or more feature vectors.
4. The method according to claim 1 , wherein, when the one or more feature vectors is determined to be ineffective, the method further comprises generating one or more proxy feature vectors for the one or more ineffective feature vectors.
5. An apparatus, comprising:
at least one processor; and
at least one memory comprising computer program code,
the at least one memory and the computer program code are configured, with the at least one processor to cause the apparatus at least to
receive raw input from an input source;
extract one or more feature vectors from the raw input;
determine an effectiveness of the one or more feature vectors based at least in part on
computing a correlation score for the one or more feature vectors,
checking the computed correlation score against a predetermined threshold, and
a signal noise level,
wherein when the correlation score is above the threshold, the one or more feature vectors is ineffective, and
wherein when the correlation score is below the threshold, the one or more feature vectors is effective;
perform, based on the determination, multiplicative fusion processing on the one or more feature vectors; and
predict, based on results of the multiplicative fusion processing, one or more emotions of the input source.
6. The apparatus according to claim 5 , wherein the raw input comprises one or more modalities.
7. The apparatus according to claim 5 , wherein, in the multiplicative fusion processing, the at least one memory and the computer program code are further configured, with the at least one processor to cause the apparatus at least to:
combine the one or more feature vectors with another one or more feature vectors;
boost one or more of the one or more feature vectors; and
suppress one or more of the one or more feature vectors.
8. The apparatus according to claim 5 , wherein, when the one or more feature vectors is determined to be ineffective, the one or more feature vectors, the at least one memory, and the computer program code are further configured, with the at least one processor, to cause the apparatus at least to generate one or more proxy feature vectors for the one or more ineffective feature vectors.
9. A computer program, embodied on a non-transitory computer readable medium and executable by a processor, wherein, the computer program, when executed by the processor, causes the processor to:
receive raw input from an input source;
extract one or more feature vectors from the raw input;
determine an effectiveness of the one or more feature vectors based at least in part on
computing a correlation score for the one or more feature vectors,
checking the computed correlation score against a predetermined threshold, and
a signal noise level,
wherein when the correlation score is above the threshold, the one or more feature vectors is ineffective, and
wherein when the correlation score is below the threshold, the one or more feature vectors is effective;
perform, based on the determination, multiplicative fusion processing on the one or more feature vectors; and
predict, based on results of the multiplicative fusion processing, one or more emotions of the input source.
10. The computer program according to claim 9 , wherein the raw input comprises one or more modalities.
11. The computer program according to claim 9 , wherein in the multiplicative fusion processing, the processor is further caused to:
combine the one or more feature vectors with another one or more feature vectors;
boost one or more of the one or more feature vectors; and
suppress one or more of the one or more feature vectors.
12. The computer program according to claim 9 , wherein when the one or more feature vectors is determined to be ineffective, the processor is further caused to generate one or more proxy feature vectors for the one or more ineffective feature vectors.