Publication

Abstract : Unlike spoken language, sign language is a representation of the culture and values of the deaf community that is easily misinterpreted or improperly translated, creating barriers in information access when there are real-time environments dominated by spoken language. This work covers the construction of a machine basedon deep learning that would be used to translate the sign language into text. The motivation for this invention is to make sign language analogous to ordinary language and close the communication gap that it creates. The data set collection covers the range of sign language videos from different people, and the preprocessing technique is included. The features are captured and based on these features a neural network model is designed that maps a sign language gesture into corresponding texts. The effectiveness of the model is seen through its data and evidence which includes accuracy. The last part of the process covers the idea for the deployment of the model in real-world situations and the commitment to a process of iterative refinement through user feedback and ongoing improvement. The results of the work are revolved around the formation of the new technologies for this community, consequently helping these people to be more integrated into communication environment.