Vocalization
What barks, whines and growls carry beyond volume: pitch, rhythm, timing.
We study canine emotion across voice, body and physiology. What we learn goes into PEBI, and what we can share is published here, including where the model still falls short.
WHAT WE STUDY
Dogs communicate with their whole body at once. Our research looks at each channel and, more importantly, at how they combine.
What barks, whines and growls carry beyond volume: pitch, rhythm, timing.
Ears, tail, posture and gaze, read together rather than one cue at a time.
How heart rate, breathing and rest add context to what a dog expresses.
Whether a read holds up across breed, size and age, and where it doesn't.
HOW WE WORK
Useful models begin with careful data. Every example has a clear origin, and every label is open to another look.
Families share voice, video and context with clear permission for research use.
Behavior and veterinary specialists label what a moment shows, and disagreements get a second look.
Every model is tested on dogs it has never seen, with results broken out by breed group, size and age.
THE EMOTION MODEL
Affective science describes emotion on two dimensions: how pleasant it feels and how activated the body is. PEBI adapts that framework to dogs, using vocal, postural and physiological markers in place of self-report.
Within each quadrant the model resolves finer-grained states, defined with behavior experts and validated against reviewed footage.
GET INVOLVED
Start with the model overview, or reach out and we’ll happily share more of the work behind it.
See what PEBI does, how it reads a dog, and what it's built for.
TAKE A LOOK→Building with Traini? We’ll walk you through the architecture, results and how we care for data.
REQUEST THE BRIEF→Universities, vets and behavior researchers: bring a question and let's explore it together.
START A CONVERSATION→