Platform
Our Models
Capture nuances in expressive behavior with newfound precision: subtle facial movements that express love or admiration, cringes of empathic pain, laughter tinged with awkwardness, sighs of relief.
Model
Face + Body
Differentiate 37 kinds of facial movement that are recognized as conveying distinct emotional meanings—and the multiplicity of ways that they can be blended together
Model
Voice
Differentiate 28 kinds of human vocalization that are recognized as conveying distinct emotional meanings—and the multiplicity of ways that they can be blended together
Model
Voice
Differentiate 18 patterns of tune, rhythm, and timbre that interact with speech to imbue our everyday conversations with complex, blended emotional meanings
Our Datasets
With hundreds of thousands of fully-consented samples, our datasets are emotionally rich, naturalistic, culturally diverse, equitable, unbiased. They are the tools needed to train and evaluate unbiased and empathic AI.
Dataset
Voice
Diverse vocal expressions (e.g., laughs, sighs, screams) from around the world, conveying blends of over 28 different kinds of emotion, with self-report and demographics
Dataset
Face + Body
Reactions to thousands of evocative videos and images by people on 4 continents, capturing dozens of distinct kinds of emotional experience, with self-report and demographics
Dataset
Face + Body
Hundreds of thousands of richly diverse facial expressions from around the world, conveying blends of 37 different kinds of emotion, with self-report and demographics
Dataset
Voice
Hundreds of thousands of recordings of sentences being spoken in dozens of different emotional intonations by people on 4 continents, with self-report and demographics
By using this website, you agree to our use of cookies.