Transcription of American Sign Language (ASL) Using Convolutional Neural Networks (CNNs)

Sign language is most natural and the fastest way to gather knowledge and information, and it is an expressive and efficient way for the people with hearing and speech impairment to convey their thoughts and feelings. Sign language to text conversion technology allows these individuals to more easily communicate with others by converting their hand gestures into text. This technology can bridge the communication gap and improve integration and accessibility for people who rely and primarily utilize sign language as their prior means of expression. This translation is done using image processing, computer vision and machine learning technique. This technology can process video input of American hand signs and convert them into text in more than one language, providing real-time captioning and subtitle improving communication and accessibility for people with HSD and with our method we get 95.7% accuracy for 26 letters. The technology not only helps people with hearing and speech impairments, but also helps in teaching sign language to those who are not familiar with it.