Reciprocal-Attention
This is a Human-Robotic Interaction project, build for QTRobot (www.luxai.com) to increase perceived engagement and to measure attentiveness especially when engaging with children.
core application - engagement tracking for chilren
recently prototyped this project with Dr. Shruti Chandra (UNBC)
overview
in this project we worked on the QT robot's main selling point which is engagement, and to improve it we created an interactive eye mechanism that utilizes facial landmark detection (Computer Vision) that responds to your gaze, using this simple prototype we could:
measure what factors affect engagement
understand when and why children attention is lost
identify ways to improve interaction rates in HRI
how it works
python tracks your face using computer vision
checks if you're making eye contact by detecting your eyes
sends the data via osc (open sound control) to a processing visualization
processing draws a digital eye that either:
tracks your movement (when you're looking at it)
closes and ignores you (when you look away)
tech stack
python 3 — face & eye detection
opencv — camera input and image processing
yolo v8 — face detection model
haarcascade — eye detection
osc (open sound control) — data transmission
processing — real-time visualization
core application - engagement tracking for chilren
recently prototyped this project with Dr. Shruti Chandra (UNBC)
overview
in this project we worked on the QT robot's main selling point which is engagement, and to improve it we created an interactive eye mechanism that utilizes facial landmark detection (Computer Vision) that responds to your gaze, using this simple prototype we could:
measure what factors affect engagement
understand when and why children attention is lost
identify ways to improve interaction rates in HRI
how it works
python tracks your face using computer vision
checks if you're making eye contact by detecting your eyes
sends the data via osc (open sound control) to a processing visualization
processing draws a digital eye that either:
tracks your movement (when you're looking at it)
closes and ignores you (when you look away)
tech stack
python 3 — face & eye detection
opencv — camera input and image processing
yolo v8 — face detection model
haarcascade — eye detection
osc (open sound control) — data transmission
processing — real-time visualization
core application - engagement tracking for chilren
recently prototyped this project with Dr. Shruti Chandra (UNBC)
overview
in this project we worked on the QT robot's main selling point which is engagement, and to improve it we created an interactive eye mechanism that utilizes facial landmark detection (Computer Vision) that responds to your gaze, using this simple prototype we could:
measure what factors affect engagement
understand when and why children attention is lost
identify ways to improve interaction rates in HRI
how it works
python tracks your face using computer vision
checks if you're making eye contact by detecting your eyes
sends the data via osc (open sound control) to a processing visualization
processing draws a digital eye that either:
tracks your movement (when you're looking at it)
closes and ignores you (when you look away)
tech stack
python 3 — face & eye detection
opencv — camera input and image processing
yolo v8 — face detection model
haarcascade — eye detection
osc (open sound control) — data transmission
processing — real-time visualization