• Mothusi - AI Powered Assistive Smart Assistance
  • An AI-Powered Assistive Smart Assistant designed to help people with disabilities by recognizing objects in their environment andresponding to voice commands with real-time voice feedback
• Installation Instructions
• Usage Instructions
• Technologies Used
  • Hardware
  • Software
• Milestones
  • Milestone 1
  • Summary
• Credits
• Known Issues
• Contact Information

• install python (make sure pip is there else install it using "sudo apt install python3-pip ")
• install utralytics (pip install untralytics)

1. Assuming that a camera (IP camera), an embedded device (raspberry pi) and a monitor (We used a laptop monitor for this project) are present.

Requirement: - AI Model for Object Recognition - Embedded System Device - Camera

• Raspberry PI 4
• Android Phone (For Camera)
• Video Capturer Adapter

• OBS Studio (To use laptop screen as a monitor)
• ultralytics (YOLOv11)# Milestones

Implemention of a basic object recognition system using a pretrained YOLOv11 model on a raspberry pi 4 using a an IP camera (From an android phone). The result of the detection are displayed on a desktop monitor.

This project is part of the coursework in CS5401 - Special Topics in Computer Science

At the moment, there is significant lag between the time a video is captured by the camera and and the output is rendered on the screen (With the bounded boxes showing a classified object an the accuracy)

• Lehooana Morie
• Tebello Ramothoto
• Lipholo Nyapa