INSA Lyon Electrical Engineerng Department : Thomas GRENIER
formation 4GE PRT : Soufian Targaoui, Yacine Youcef
Centre Blaise Pascal : Emmanuel Quémener

Many deaf people use sign language to communicate. This language involves hand gestures, body movements and facial expressions to convey the meaning of words and sentences.

In France, for a long time, sign language was repressed and therefore not promoted. Recently, its use has evolved in a positive way but, for the French language, many words are devoid of signs and more particularly the words of modern science. A second impact of this policy is that few people who are not deaf or dumb understand this language.

Deep learning is an effective approach to the recognition of images and gestures. Recent studies have proposed architectures dedicated to the recognition of sign language that give interesting results. In this project, we propose to automatically recognize 10 French signs dedicated to the science of electrical engineering using a deep learning approach.

The Centre Blaise Pascal provides :

• support and ressources for deep learning training and optimization
• expertise in GPU programming