While India continues to make great progress on the economic front, the country also has a disturbing statistic. With close to 52 million visually impaired people, India has one of the largest populations of the visually impaired. Researchers say that this is close to one fourth of the world’s visually impaired population.
This compelling issue has inspired Persistent Systems to take this challenge head on in its internal hackathon, Semicolons. It is a 24-hour hackathon where programming teams compete to solve complex problems by exploiting the latest technology. During the brainstorming session, one of the earliest ideas that came up to the team was to do something around accessibility technology and the team eventually settled on making the world more accessible to the visually impaired.
Accordingly, the team developed Drishti – a mobile app that makes the visual world accessible to the blind by describing it to them. Explains Pandurang Kamat, Chief Architect – Innovation and R&D, Persistent Systems, “Over 280 million people worldwide are visually impaired and around 40 million are completely blind. We feel a technology like Drishti will render the world more accessible to them. Drishti uses statistical learning using Naïve Bayes model in Scikit-learn to identify and classify the scene. The model learns from hundreds of images which have been classified and tagged already – supervised learning. It also uses a deep neural network model to recognize faces using Facenet toolkit. For this, we fed the tagged photos of a user’s friends taken from their Google photo albums. It uses multiple computer vision APIs to identify objects and describes the high-level scene in English using a NLP grammar module.”
Persistent Systems’ efforts in applying AI for a social impact is laudatory, and if successful, can transform the lives of millions of visually impaired people.
If you have an interesting article / experience / case study to share, please get in touch with us at [email protected]