Icoobook
An interactive picture book that reads emotion back to the child reading it.
2023 — 2024
Sep 2023 — Mar 2024
Student Team Lead — Product design, technical direction, user research
Figma · Python (DNN) · Word2Vec · Adobe Illustrator
#EdTech · #Children · #Affective Computing
A children's reading platform that recognizes the emotion in a child's voice as they recite poetry, then re-illustrates the scene to match. A collaboration with Tsinghua University, showcased at ACM Multimedia.
The brief
Children’s reading apps tend toward two extremes: gamified to the point of distraction, or so passive that engagement collapses. The team set out to find a middle path — something that meets a child where they are, emotionally, and quietly invites them deeper.
I led the team as student PI, in collaboration with Tsinghua University.
The system
Three interlocking modules, each tuned to a different cognitive register:
- Linking Strokes — a reading-comprehension game where children connect strokes to advance the story.
- Magic Brush — children draw an answer to a prompt; the system converts the sketch into a story-world object.
- Reading Poetry — the centerpiece. The child reads aloud; a deep-neural-network model recognizes the emotional tone (joyful, melancholy, neutral, brave) from acoustic features (F0, energy, MFCC) and word embeddings (Word2Vec). The illustration that accompanies the verse re-paints itself to match the emotion.
The result is a story that listens.
Research & validation
I designed and ran an experience study with ~40 children, comparing Icoobook against a control reading app. Across attention, cognitive memory, and emotional satisfaction, Icoobook outperformed the baseline at statistically meaningful levels.
Recognition
The project was invited for showcase at ACM Multimedia, a top venue for multimedia computing research.
My contribution
End-to-end product design, interaction architecture, and the technical-research path that bridged the design team and the AI team at Tsinghua. Coordinated user research, ran the experience study, and shaped the final visual language for a child-friendly UI.