- Ibrahim Sobh
- Patrick Saade
- Isaac guillermo Gonzales
While it is easy to overestimate the effectiveness of communication. The results of poor communication skills can keep a child from progressing at school and lead to stress within the family. That is why our research aims to strengthen the bridge of communication between parents and children in the Deaf community using a few-shot sign language recognition model. Our goal is to break the sign language learning barrier that most families struggle to pass through, by providing an effective layer of communication that can work as the intermediate translator between the child and his parent.
The challenge every data scientist faces nowadays is having little to no data to work with. Labelling and pre-processing additional samples is a time-consuming and expensive task to deal with. Even if data augmentation delivers better results, it is not always the key to solving this problem. We need to innovate new methods to deal with the lack of information in some domains or fields where it is crucial to identify the problem at an early stage. One of the hottest topics in the field related to solving this problem is Few shots, one shot learning models, this kind of model comes in handy when it comes to reducing data collection and computational costs as well as creating a test base for learning like a human, a bright new aspect of machine learning where computers will be learning from a few examples like humans. This will revolutionize the way we perceive data science.