Inspiration
Anywhere we see, deaf people have a hard time understanding and comprehending the language that everyone speaks. This is due to the fact that they can’t listen to what the other person wants to say, thereby getting confused and disoriented on a day to day basis. Sign language helps build communication in addition to bridging missing gaps. When you learn sign language and practice it, it can actually help you with all aspects of communication; even children who aren't hearing impaired can benefit from it if they use it during their early development years
What it does
.We will be serving the Deaf Community all over the world. Our App name is “HT-Lang” It will convert our speech into text and our texts into speech. Additionally, our text or speech are translated into the Sign Language that are displayed visually.
How I built it
We propose an efficient cascaded model for sign language recognition taking advantage from Spatio-temporal hand-based information using deep learning approaches, especially Single Shot Detector (SSD), Convolutional Neural Network (CNN), and Long Short Term Memory (LSTM), from videos. Technology used in the proposal: It is just a concept, which we intend to demonstrate using Python3,since it is very user-friendly with its usage and has a lot of pre-built libraries and packages that can be used effectively. Using libraries like Keras , Tensor Flow, OpenCV ,we intend to make the model such that it has a high accuracy of understanding the speech or text . Proposed Cost: Since it is merely a proposal ,for building the prototype there is no cost required. The expected cost of development and publishing the app is around $10000. For scaling the same to a larger user base and maintenance purpose(including bug fixes ,updates) we expect around 20% of initial cost of build, or about $2000.
What's next for HT-Lang
Where it can be used: This app can be used all over the world, the only requirement being a smartphone with internet connectivity. Our initial conversion is based on American Sign Language into English, but we plan to expand it to other standards ,such as British Sign Language ,Indo-Pakistani Sign Language, French Sign Language ,etc. and incorporating other languages too ,along with English, such as Hindi, French, German, Mandarin and other local languages.
Built With
- deeplearning
- language
- machine-learning
- natural
- processing
Log in or sign up for Devpost to join the conversation.