Synopsis

The amount of deaf and mute individuals on the earth is rising at an alarming rate. Bangladesh has about 2.6 million people who are unable to interact with the community using language. Hearing-impaired citizens in Bangladesh use Bangla sign language as a means of communication. However, these sign languages can be a bit difficult for typical people to understand. Hence, an intermediary step is needed for smooth communication between typical and hearing-impaired people. Hence, a deep learning based model can help in easing the process of identifying sign languages that can remove barrier between those two groups of people. This is can affect the society in a positive way and offer improvement to the day to day communication for those people in need. 

Relevance of the Topic

The topic is a very important as it directly correlates with the lifestyle improvement for the disabled people. 

Future Research/Scope

• Build a deep learning model to identify sign languages.
• Create an intermediary interface for communication between typical and hearing-impaired people.
• Integration of the model into a workable hardware/software for easier interpretation.

Skills Learned

• Basics of machine learning.
• Deep learning techniques such as CNN.
• Different CNN architectures such as AlexNet, VGGNet, ResNet, etc.
• Usage of high level neural network APIs such as keras.
• Usage of different libraries such as Numpy, Matplotlib, etc.
• Usage of different model performance metrics such as precision, recall, f1 score, confusion matrix, etc. 

Relevant courses to the topic

• Artificial Intelligence (CSE422)
• Neural Networks (CSE425)
• Machine Learning (CSE427)
• Image Processing (CSE428)

Reading List

• Huang, J., Zhou, W., Li, H. and Li, W., 2015, June. Sign language recognition using 3d convolutional neural networks. In 2015 IEEE international conference on multimedia and expo (ICME) (pp. 1-6). IEEE.
• Pigou, L., Dieleman, S., Kindermans, P.J. and Schrauwen, B., 2015. Sign language recognition using convolutional neural networks. In Computer Vision-ECCV 2014 Workshops: Zurich, Switzerland, September 6-7 and 12, 2014, Proceedings, Part I 13 (pp. 572-578). Springer International Publishing.
• Hossen, M.A., Govindaiah, A., Sultana, S. and Bhuiyan, A., 2018, June. Bengali sign language recognition using deep convolutional neural network. In 2018 joint 7th international conference on informatics, electronics & vision (iciev) and 2018 2nd international conference on imaging, vision & pattern recognition (icIVPR) (pp. 369-373). IEEE.
• Hoque, O.B., Jubair, M.I., Islam, M.S., Akash, A.F. and Paulson, A.S., 2018, December. Real time bangladeshi sign language detection using faster r-cnn. In 2018 international conference on innovation in engineering and technology (ICIET) (pp. 1-6). IEEE.
• Cui, R., Liu, H. and Zhang, C., 2017. Recurrent convolutional neural networks for continuous sign language recognition by staged optimization. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 7361-7369).