Coin classification
September 27th, 2010
There are thousands of coins sitting in large safes around the world waiting to be catalogued. The task of cataloguing is a painful process of looking up old and damaged coins in relatively old clumsy books and then manually retyping the information from the book into the institutions own catalogue system.
My current focus is in classifying coins based on the images that appear on the obverse/reverse of the coins. This is a multi-class problem in the scale of 30-50 classes. This coin dataset is extremely difficult due to odd coin shapes, relatively low contrast of details, and a lot of rust damage. Due to these characteristics methods such as Surft, Sift, and SVM classification schemes fail very badly.
The current strategy for the coin dataset is in using multiple “expert” hierarchical classifiers to manage this large multi-class classification problem.