A robot that draws paintings onto a blind person's palm with warm and cold brushes, with narration.
Problem
Galleries are mostly 2D visual work; blind and low-vision visitors cannot experience paintings.
Solution
Segment a painting into meaningful objects, convert each into simplified stroke graphs and millimetre coordinates, and have a two-axis robot trace them on the palm; warm vs cold colours map to heated vs cooled brushes, with narration and mood music in a Bob Ross voice.
Demo moment
Photograph a painting by voice, then feel its shapes drawn on your hand while a familiar voice describes them.
Why it stood out
- curator inferenceClear human stakes: a specific group gets access to culture they are excluded from.
- team statedThe warm/cold two-brush feature was not working; the team demoed one brush and no colour, and still reached the finals — a working core beat a complete feature list.Source ↗
- team statedDropped an LLM-based simplification approach for a deterministic geometric pipeline after inconsistent results.Source ↗
- curator inferenceBob Ross as name and narrator makes the accessibility message warm and memorable.
- organizer statedFeatured by the organizer as a past showcase project.Source ↗
Lessons for your next hackathon
- Cut the broken feature from the demo and describe it as next; never let the unfinished part define the pitch.
- When an LLM step is flaky, replace it with deterministic code before demo day.
Projects that share domains, mechanisms or judging lenses — often with a different problem and stack.Tune the comparison.
Award winner · 2025
A haptic vest and Braille gauntlet that guide DeafBlind people to objects and help them communicate.
Shared: domains accessibility, robotics hardware · mechanisms physical artifact, human stakes, visible engineering depth · lenses impact, technical depth · ai computer vision, voice ai · stack python
Finalist · 2025
A real chessboard whose pieces move by themselves, so a remote player can face someone at the board, with coaching and voice control.
Shared: domains robotics hardware, accessibility · mechanisms physical artifact, human stakes, visible engineering depth, demo first wow · lenses wow factor, technical depth, impact · ai computer vision, voice ai · stack arduino, python
Award winner · 2025
Smart glasses plus a dynamic Braille display that describe surroundings and people to visually impaired users.
Shared: domains accessibility, robotics hardware · mechanisms physical artifact, human stakes, visible engineering depth · lenses impact · ai computer vision, voice ai · stack arduino, python
Grand winner · 2025
A turret that finds each plant with a camera and shoots it exactly the water it needs, controlled from a web app.
Shared: domains robotics hardware · mechanisms physical artifact, demo first wow, visible engineering depth · lenses technical depth, wow factor, impact · ai computer vision · stack arduino, servos, python
Award winner · 2026
A kagent-based agent that captures webcam photos and transforms them into AI-augmented media, with an American Sign Language mode.
Shared: domains creative media, accessibility · mechanisms demo first wow, human stakes · lenses wow factor, impact · ai computer vision, generative media
Podium #3 · 2025
Smart gloves that turn walking directions into vibration patterns, with a voice agent for planning and siren alerts.
Shared: domains accessibility, robotics hardware · mechanisms physical artifact, human stakes, visible engineering depth · lenses impact, technical depth · ai voice ai