Accessibility Without Borders

Bringing Inclusivity in Tech

Comment

Inspiration

Accessibility Without Borders is a hack for social good. The project aims to give people the power and reduce the barriers. The goal is to show how apps can be inclusive and accessible.

What it does

There are 3 tools in this project:

  • Summarization – A summary tool that converts long texts into short digestible texts featuring only the most important information. This tool aims to help people with ADD, ADHD, dyslexia, and those who are lower literacy.
  • Emotion Detection– A tool that classifies emotions from text input. This tool aims to help people with autism identify emotions and help them connect and communicate.
  • ASL Detection - A tool that helps to break barriers for people with auditory disabilities by converting ASL to speech in real-time using AI

How I built it

  • Summarization - Using TextRank in NLP to get a summarization
  • Emotion Detection - Took data from 4 datasets and then did the following: preprocessed, feature engineering, data cleaning, transformations (TFIDF). Then made multiple models using a RobustScalar on the data.

  • ASL Detection - Used Mediapipe to generate points on hands, then use those points to get training data set. I used Jupyter Notebook to run OpenCV and Mediapipe. Upon running our data in Mediapipe, we were able to get a skeleton map of the body with 22 points for each hand. These points can be mapped in 3-dimension as it contains X, Y, and Z axis. We processed these features (22 points x 3) by saving them into a spreadsheet. Then we divided the spreadsheet into training and testing data. Using the training set, we were able to create 6 Machine learning models:

  • Gradient Boost Classifier

  • XGBoost Classifier

  • Support Vector Machine

  • Logistic Regression

  • Ridge Classifier

  • Random Forest Classifier

Challenges I ran into

  • Teammate quit had to work solo.
  • 24 hours is not enough time, had to ditch two more tools planned for this project
  • Couldn’t work on Front-End due to a lack of time and experience

Accomplishments that I’m proud of

Made a project successfully that can help people and bring communities closer.

What I learned

  • Time management! 24 HOURS!
  • TextRank (NLP) was very challenging but rewarding.
  • Project management
  • Enhanced Mediapipe/OpenCV

What's next for Accessibility Without Borders

  • Collect more data to do more classifications (for Emotion and ASL detection)
  • Build a polished app
  • Sell the backend (ML models) to companies to incorporate accessibility options easily in their services.
  • Make more tools such as helping people with visual impairment.

Built With

Share this project:

Updates