Sign Language Recognition using CNN

• A real-time gesture recognition is an important factor in facilitating communication between the hearing impaired community and the common public.
• The proposed method for developing a Sign Language fingerspelling translator is based on convolution neural network (CNN).
• We have utilized OpenCv to process the images and obtain the region of interest(ROI) for the CNN model to recognize the sign.
• We produced a robust CNN model using Keras API and Tensorflow-Python Library that consistently classifies letters a-z correctly in a majority of cases.
• The resultant accuracy of 81.39% fulfills our purpose to prepare an intelligent system which can work as a translator between the sign language and the spoken language in real time.
• Most alphabets could be recognized correctly and average recognition time was also less.