As I had 2 weeks of classes in Research Methods, now MS Thesis topic is being shaped in my mind. Me and my colleague Ibrahim was thinking about how to develop a system that will ease the lives of disabled people of the community from the first day. It is fact that there are several communication obstacles in people who cannot talk with the rest. This is where idea came off.
Our MS Thesis topic is developing Sign Recognition system which translate the video in sign language to the text. As most of the countries have their own and different sign language, our domain is Azerbaijanian language. It is both Computer Vision and Machine Learning project as we will get the features from video and train those data to get corresponding text. There is also Natural Language Processing task that from a particular video we will get sequence of words. Generating meaningful sentences from those words is NLP task.
However, it is seen that we would have some problems about finding appropriate data. In this manner, we will contact governmental officials to help us with the data. In addition, I believe we can find some data online such as in Kaggle. We will surely need the help of trainers who are the professionals in sign language of Azerbaijanian.
In this short blog, I have explained our MS Thesis topic and the data we are willing to use. Right now, we are in the process of researching. We read different research papers on this topic and look through its applications. We have not decided the techniques and models we will use yet, but for now, CNN + LSTM seem best option for now. We plan that till the end of ‘Research Methods’ class, we will have clear plan that we will follow till the May 2022.
This is a great start. I've left comments on your team-mate's blog. You have a strong plan for your project, so I think that getting started on any data you can get access to would be a great next step!