Docket #: S19-296
Automated Recognition of Facial Expressions with Neural Networks
Artificial intelligence can be leveraged to evaluate how facial expressions will be perceived by others. A deep learning neural network is used to generate facial vectors for each image of a person. Associated traits, including attractiveness, competence, etc., are rated by participating individuals. A model is trained to analyze each image's deviation from baseline traits, or the average of all other photos, based on facial vectors. This technology can be useful for helping individuals select appropriate photos for different online settings, as well as observe and assess "first impressions" in real-time.
Applications
- Picture selection for social and professional websites
- Real-time feedback for how photos or streamed video may be perceived
Advantages
- Robust and accurate predictions of perceived facial expressions
- Multiple dimensions underlying facial judgement evaluations
- Scores images relative to other photos of the individual, not among other people
- Focus is on within-person variation, not between-person variation
- Still images or real-time video can be analyzed
Similar Technologies
-
Trainable Analog Hardware Platform for Recurrent Neural Networks (RNN) S19-165Trainable Analog Hardware Platform for Recurrent Neural Networks (RNN)
-
End-to-end Semantic Segmentation of 3D Point Clouds by Convolutional Neural Networks, Trilinear interpolation, and Conditional Random Fields S17-380End-to-end Semantic Segmentation of 3D Point Clouds by Convolutional Neural Networks, Trilinear interpolation, and Conditional Random Fields
-
Denoising WaveY-Net: An ultra-fast, auxiliary neural network enhanced surrogate field solver S22-445Denoising WaveY-Net: An ultra-fast, auxiliary neural network enhanced surrogate field solver