boot camp day 3

this morning i had the classic 'everything goes wrong' type of dream... the type of thing where i was on a bus going to the airport, but it went to the wrong place, so i had to find my way to the airport, and i went to the ticket counter, but they couldn't find me on the computer, so i went to check my phone to see the email with the record locator, but my phone had been reset and i couldn't access anything... so then i had to answer some security questions on a screen to gain access to my ticket information, but of course the questions were incredibly cryptic, like riddles, or they were just symbols that i had to interpret, etc. etc. etc.

boot camp went well today. in analysis, we got through the arzela ascoli theorem. in linear algebra, we got through dual spaces. during lunch, i joined a big group of first year students and walked over and sat outside at the union and ate lunch, which was just fine and dandy.

after classes were over, i went and saw a talk by Jean-Michel Morel about the "retinex algorithm." the topic had to do with how we perceive images... probably everyone has seen this famous optical illusion. the point is that there are two parts of the image that are exactly the same color, but because our brains are smart and can filter out shading, we perceive the two parts of the image as being two different colors. so, people have tried to come up with algorithms that take an image and filter out the shading, so that the resulting image is closer to what we actually see... e.g., if you perform the retinex algorithm to that image, the part that looks darker actually _is_ darker and the part that looks lighter actually _is_ lighter. basically, there are 3 different models of how this works, and when you use these models to create algorithms to reconstruct an image with the shading removed, you basically end up having to solve the same partial differential equation (PDE) in each case... it's not exactly the same, but it's close. anyway, a lot of the talk was over my head, but i did learn some things, like the idea of the discrete laplacian being an operator used for 'edge detection' (even though i don't know how this works! but i would love to, if anyone could tell me) and how you can basically get rid of gradients below/above certain levels to yield an image with different levels of contrast. anyway, it was neat and i'm glad i went. here is a link he provided during the talk to see some of this stuff in action.

after that talk, i met with my faculty advisor to plan classes for the year... in the fall i'm going to take complex analysis I, real analysis I, and applied PDE I. (plus the class that's supposed to teach you how to be a TA.) technically, i am taking classes that an applied math PhD student should be taking! but that's okay... i'll do the geometry/topology sequence next year (schedule-wise, it's not possible for me to take it now)... also, in the winter and spring, terry tao is giving a topics in real analysis course which i am thinking i will take if the topics aren't too above my head. we also charted out a good potential bike route to the beach that involves bike paths and avoids deadly streets...

anyway, cardinals game is on, and i am thinking i would like to get some work done now even though i feel that a nap may be more likely in my near future... maybe a music blog post, but that might have to wait til friday... anyway, thanks for reading!