Tuesday, January 10, 2006

Catching up...


I started work on monday. My primary contact was out sick. I instead spoke with my secondary contact. She showed me to the workstation I would be using and introduced me to a few people who work there. Unfortunately, my workstation did not yet have internet access. I'm assuming that the workstation has not been used for a while and has been disconnected. She was nice enough to download Louka's thesis on car recognition to a USB drive which I then transferred to my laptop. I spent the next 3.5 hours or so reading over Louka's thesis, taking notes on it, and coming up with questions to ask my professor about. Here are some vocabulary words of interest, placed in the context of Louka's paper:

LPR: License Plate Recognition

MMR: Make and Model Recognition (referring to the make and model of cars)

Strong Classifier: a window that is placed over all possible locations in a given frame. The strong classifier returns true if the frame contains a license plate and false otherwise. Whether the strong classifier returns true or false is determined by weighing the decisions of many weak classifiers.

Weak Classifier: Given the contents of a strong classifier, a weak classifier detects the presence of a single feature. The weak classifier returns true if the feature is detected and false otherwise.

Feature: A feature is a single detail in an image, such as the x-derivative or y-derivative, that aids in the detection process.


Post a Comment

<< Home