Objectives: The sign language is very important for hearing impiared people. Finding an educated translator for the sign language every time and everywhere is difficult task. The human-computer interaction system is helpful for dumb people to overcome the difficulty, besides it and can be installed anywhere. This paper proposes the method or algorithm for an application which would help in recognizing the different signs and convert those sign gestures into voice. Methods: Different sets of hand gestures were captured using web camera and then stored in a directory. The correct signs by the user is identified by using feature extraction techniques and neural network algorithm. Findings: The sign languages for different numbers in words are trained and tested. The test image is aligned correctly with training images which is based on correlation and convert the matched image into text and then text into voice. Applications: By using this system, hearing impaired people can easily interact without depending on translators.
展开▼
