I have a master degree in image processing (including a two-year face recognition capstone project).
If I find myself upvoting posts in other areas of signal processing, I might be wrong - because I lack the knowledge to detect garbage in unfamiliar areas.
My mathematical understanding is not strong - I tend to find ways to visualize the problem and the mathematical tools. As such, I have poor understanding of things which can have infitinely many dimensions. I understand PCA, LDA (linear discriminant) as being useful for dimension reduction, though.
I'm working for a document image processing company, with 5-10% research, >90% general C++/MFC/Windows programming. I am an avid user of MATLAB, though it is only used internally and not used for production software.
Most of my results and work is ad-hoc (specific to an application, and not deeply rooted in some profound mathematics). So far, I have not made any "contribution" (new theories) in the field of signal processing.