By Amanda Weissman
It’s one thing for engineers to implement a solution for a community, but it’s another thing to create a solution hand-in-hand with a community. The students from the IEEE CIS Student Branch of M.S. Ramaiah Institute of Technology, understood this as they began their EPICS in IEEE project.
This student team has taken on the challenge of creating an affordable, accessible, and open-source device that is tailored to the wants and needs of the blind community. The device will assist in everyday tasks, such as reading aloud written text and notifying the user how far away a specific object is.
Working with a Community Partner
This EPICS in IEEE-funded project began after the students saw firsthand how limited the visually impaired options are due to both expensive devices and inaccessible braille education. “We saw that one of the devices, Orcam, was about $3,000. So it’s quite expensive and inaccessible, especially for people in India,” says Tejas S, one of the student team members. The student team is focussing on creating not only a more affordable device but also one that is comfortable and friendly to the blind user.
The student team is in continuous contact with the National Association for the Blind, Bangalore, and is working collaboratively with some of the blind students to ensure their product is usable. “Merely talking and working with the National Association for the Blind helped us look at problems from a different angle,” Tejas MR says.
For example, Looking at the physical features from another angle made the student team realize they should use bigger buttons with a specific texture. During a recent visit to the association, the team was fortunate enough to get direct feedback like this from the blind students. The student team asked the participants questions pertaining to the device’s likeability and functionality.
Utilizing Open Source
After a successful round of feedback on their hardware and imaging, the team began developing the software portion, which is to be completed via open source. Open source allows the team to study, change, and use the software and source code that have granted permission to be used by others.
“To create an OCR model from scratch, it would take millions of images, complicated architecture, and more, so having models like tesseract, created by Motorola, is very helpful for projects like these,” Tejas S explains. Essentially, by utilizing open source they can work in a collaborative manner to achieve their goals faster and also learn along the way.
“If we don’t develop the whole part, someone else can get a hold and look at it later on,” Tejas S says. With the largescale goal of selling their device to a company that can mass produce and distribute them, it’s beneficial that the model’s open-source software can be further developed, modified, and improved.
The device will have a basic model that works offline with open source, and a premium model that uses Google API, which costs a monthly subscription. While the tesseract model is 90-95% accurate compared to Google API, not everyone can afford the recurring payments of the Google model, which is why the team decided to create a $200USD basic, offline open source model. The basic model can scan any text and translate it to the user’s regional language via speaker or headphones, as well as relay how far an object is from the user, such as a pole in their walking path.
EPICS in IEEE has given the student team an opportunity to access needed resources to build the prototype and resources to engage with the community. Throughout the project, the students have improved their problem solving, budget managing, and hands-on engineering skills.