[zhivota]

Just as an alternative viewpoint, I worked at a DOE national laboratory on a big image processing system for the USAF, and we used C++, Python, and the PIL. Believe it or not PIL is really fast, in fact I tried to optimize the C and couldn't do much with it, really. (Of course that probably says more about me than anything!)

Just saying that Python has massive inroads in scientific computing, so you can't go wrong learning it for that purpose. Just make sure you keep your eye on C/C++ as well, as you will need it to do your mathematical algorithms - Python libraries are fast because they are coded in C, so if you do anything computationally heavy that can't be offloaded to a library, it should also be in C.