Skip to content
HELMER
Open to buildABJ --:--:--
← Portfolio / Engineering

PERRY

An offline voice-recognition smart-home system built as assistive technology for people with motor disabilities, using CNN models trained on a manually collected Nigerian voice dataset. Presented at ICAIR-MIRG 2025.

STATUSICAIR-MIRG '25
YEAR2025
DISCIPLINEEngineering
UNDEREngineering
PERRY — hardware prototype — hero shot
Detail / mobile
Flow / screen recording
01 / The brief

Voice-controlled smart home

An offline voice-recognition smart-home system built as assistive technology for people with motor disabilities, using CNN models trained on a manually collected Nigerian voice dataset. Presented at ICAIR-MIRG 2025.

02 / What it does
  • Runs fully offline on a Raspberry Pi
  • Trained on a Nigerian-accent voice dataset
  • Presented at an international AI conference
03 / Stack
PythonCNN/LSTMMFCCRaspberry PiArduino
Ask about this build →