India (Commonwealth Union)_ In a groundbreaking development, Priyanjali Gupta, a third-year engineering student at the esteemed Vellore Institute of Technology (VIT) in India, has unveiled a remarkable achievement: an Artificial Intelligence (AI) model capable of detecting and translating American Sign Language (ASL) into English in real-time.
Priyanjali shared a captivating demo of her AI-based ASL Detector on her LinkedIn profile, showcasing its incredible potential. While the current iteration supports a selection of essential phrases such as “Hello,” “Please,” “Thanks,” “I Love You,” “Yes,” and “No,” it is an inspiring glimpse into a future where technology bridges communication gaps effortlessly.
The heart of Priyanjali’s creation lies in her adept utilization of the Tensorflow object detection API, coupled with the application of transfer learning through the pre-trained model ssd mobilenet. This ingenuity allowed her to repurpose existing code, tailoring it to craft her ASL Detector model. It’s worth noting that the model doesn’t precisely translate ASL to English; rather, it recognizes the signs, gauging their similarity against pre-programmed objects in its database.
Speaking to Interesting Engineering, Priyanjali shared her motivational journey behind this innovation. She candidly recalled how her mother’s encouragement, or rather “nagging,” to “do something” during her tenure at VIT, sparked her contemplation. This gentle nudge led her to ponder how she could leverage her burgeoning knowledge and skillset for the greater good. One day, amid a conversation with her trusty virtual assistant, Alexa, the idea of inclusive technology struck her like a bolt of lightning. That pivotal moment ignited a series of plans that would culminate in her remarkable creation.
Priyanjali also acknowledged the invaluable influence of YouTuber and data scientist Nicholas Renotte, particularly his 2020 video detailing the development of an AI-based ASL Detector. Renotte’s insights undoubtedly provided her with a crucial foundation for her own project.
While Priyanjali’s LinkedIn post garnered a flood of positive responses and accolades from the community, an astute AI-vision engineer pointed out that the transfer learning approach employed in her model is a technique “trained by other experts” and considered one of the “easiest things to do in AI.” Priyanjali took this feedback gracefully, acknowledging that crafting a deep learning model exclusively for sign detection is indeed a formidable challenge, though not an insurmountable one.
Modestly, she remarked, “Currently I’m just an amateur student, but I am learning, and I believe sooner or later our open-source community, which is much more experienced and learned than me, will find a solution. Maybe we can have deep learning models solely for sign languages.”
Curious minds eager to delve deeper into Priyanjali’s project can explore her GitHub page, where a treasure trove of resources awaits. Her dedication to making technology more inclusive and accessible is a testament to the boundless potential of young minds.
Priyanjali Gupta’s ASL Detector is not just a triumph of engineering prowess; it’s a beacon of hope for a more inclusive future. By breaking down barriers in communication, Priyanjali has demonstrated the power of technology to bring people together, regardless of linguistic differences. Her achievement is a reminder that innovation knows no bounds, and with passion and determination, we can overcome any challenge.
