Most of our work so far has been software engineering -- just getting our implementation of these algorithms to work. This work will last a little while longer, there are a few things left including GPU acceleration, real-time recognition, and pre-processing, but none of these threads are likely to lead to publications. For that, we need to do a literature review.
We need to research what has been done with PCA/LDA/ICA in the last 10 years, and look for papers that present a problem that we can address. In other words, we need to look at the "Future Work" sections of any relevant papers for things that we can work on. I will try to use this issue thread to document any papers that I find.
Most of our work so far has been software engineering -- just getting our implementation of these algorithms to work. This work will last a little while longer, there are a few things left including GPU acceleration, real-time recognition, and pre-processing, but none of these threads are likely to lead to publications. For that, we need to do a literature review.
We need to research what has been done with PCA/LDA/ICA in the last 10 years, and look for papers that present a problem that we can address. In other words, we need to look at the "Future Work" sections of any relevant papers for things that we can work on. I will try to use this issue thread to document any papers that I find.