Indian engineer develops AI model that instantly translates American Sign Language to Eng

Priyanjali Gupta, an engineering student from Tamil Nadu's Vellore Institute of Techlogy (VIT), has created an AI model which can translate American sign langu to English in real-time. A third-year computer science student, Priyanjali Gupta is specialising in Data Science and developed new model using Tensorflow object detection API and it is able to translate signs using transfer learning from a pre-trained model named ssd_mobilenet. 

Gupta has shared her creation on LinkedIn wherein she demonstrated capabilities of her AI model in a demo video. According to her Github post, " dataset is made manually by running Im Collection Python file that collects ims from your webcam for or all mentioned below signs in American Sign Langu: Hello, I Love You, Thank you, Please, Yes and ", She also displayed same in her demo clip. 

Her post has garnered almost 60,000 likes on platform as many seem to be impressed with invative idea that makes communication more smooth and effective. 

Used mor's taunt as inspiration

It was in February last year when Priyanjali Gupta got idea of developing something while pursuing her Engineering degree. In an interview with Interesting Engineering, she said, "She (mor) taunted me. But it made me contemplate what I could do with my kwledge and skillset. One fine day, amid conversations with Alexa, idea of inclusive techlogy struck me. That triggered a set of plans", she told Interesting Engineering. Ackwledging that creating a deep neural network for sign detection is a complex task, she said that she is learning and furr working on it. 

Talking about various or platforms using techlogies for translation of sign langus, she told IE, "According to me, researchers and developers are trying ir best to find a solution that can be implemented. However, I think first step would be to rmalize sign langus and or modes of communication with specially-abled and work on bridging communication gap". 

Im: LinkedIn/@Priyanjali Gupta