Skip to content

Hackathon organised by Taylors University

License

Notifications You must be signed in to change notification settings

szeyu/1st_Day_Hack

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

33 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

i-Lens 😎

Introduction

This web application implements Tensorflow to allow people who are blind or visually impair to move to other places which they haven't went before without guidance and assistance for other people.

Installation

This application is tested and build in the conda environement with python 3.8.5. To prevent compatibility issues, please install Python 3.8.5 using the command conda install python=3.8.5 Install OpenCV from the official page or through command line pip install opencv-python

For tensorflow

use this command to install everything tensorflow need:

pip install pillow Cython lxml jupyter matplotlib contextlib2 tf_slim

for more information about tensor flow can find out at https://www.mygreatlearning.com/blog/object-detection-using-tensorflow/

For PyTorch

to install Pytorch, please visit the page https://pytorch.org/ and download the correct version for you machine.

Requirements

use this command to install all the requirements:

pip install -r requirements.txt

To Run This Web App

run live object detection

use this command to run in streamlit:

Note that you have to cd inside the research file first

streamlit run realtd.py

run text detection

use this command to run in streamlit:

streamlit run text-detection.py

Note

Some of the elements / widgets will not work unless you change the directory to your local machine directory. eg. C://Document//research//images_text_detection If the error file not found exist, they consider making the directory to absolute directory instead of relative.

What it does?

  • It can help to detect object in front of the blinds and instruct him/her to go left, right, continue walking or stop.

  • It can help to detect text and convert to speech.

Current Cons

  • Text detection to speech cannot be combine together in the real time obstacle avoidance system due to streamlit limitations, currently it is store as another file. In the future, text detection to speech will be implemented in the main application.
  • Currently this application is running on streamlit and only works PC's. In future, we will implement this application to be use in small portable cameras and become the best companion for the blind :)

Demo

You can watch YouTube video link below about how our Web App work! https://www.youtube.com/watch?v=0LM8Xf1nGzM&feature=youtu.be&fbclid=IwAR1ekuggGRjHrgmBexOhr8xyrm-5AGh7GAuMy-SWC16nls788Gxza3xtMM4&ab_channel=TanCK

About

Hackathon organised by Taylors University

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 45.5%
  • C++ 37.2%
  • Makefile 8.2%
  • Jupyter Notebook 4.8%
  • Shell 2.0%
  • NASL 1.0%
  • Other 1.3%