VOICE INTERACTION
“Hello Orbi” activates the assistant, recognizes commands, processes the request, and speaks results back.
AI · ACCESSIBILITY · COMPUTER VISION
ORBI
ASSISTIVE AI FOR THE VISUALLY IMPAIRED
An AI-powered assistive platform that helps visually impaired users understand their surroundings through voice-controlled computer vision.
01 / PRODUCT SHOWCASE
An interface designed for hands-free awareness.
The live dashboard brings detection, text reading, device preferences, timestamped activity, and location into one accessible product flow.
MY ROLE
PRIMARY / SOLO DEVELOPER
FULL-STACK DEVELOPMENTUI/UX DESIGNPRODUCT IDEATIONAI INTEGRATIONCOMPUTER VISIONSYSTEM WORKFLOW
I handled product ideation, user flow, UI/UX, frontend and backend implementation, computer-vision integration, AI features, and feature planning.
03 / CORE CAPABILITIES
Designed around independent interaction.
OBJECT DETECTION
OpenCV captures frames while YOLOv8 identifies objects and estimates distance, such as Apple · 30 cm.
TEXT RECOGNITION / OCR
EasyOCR scans captured text, cleans recognition output, and forwards it to text-to-speech.
LOCATION & TIMESTAMPS
Location is shown on a map while activity timestamps provide useful context for navigation and tracking.
04 / VOICE-TO-VISION FLOW
A simple flow for a complex task.
01VOICE INPUT
02ORBI ACTIVATED
03COMMAND RECOGNITION
04AI PROCESSING
05VOICE OUTPUT
USER↓VOICE / CAMERA INPUT↓FRONTEND + BACKEND↓YOLOv8 · EasyOCR · OpenCV · SPEECH↓TEXT / VOICE / MAP OUTPUT
05 / TECH STACK
Technology for accessible perception.
AI / COMPUTER VISION
YOLOv8 · Roboflow · OpenCV · EasyOCR
SPEECH
Speech Recognition · pyttsx3
WEB / DEVELOPMENT
Python · Frontend · Backend · Google Maps
FUTURE WORK / 01
NEO-6M GPS MODULE
A dedicated GPS module can support more accurate location tracking and navigation in future iterations.
Explore the Orbi Assistive AI Repository
View the implementation behind Orbi’s voice-controlled computer vision, assistive OCR, and navigation assistant.