We investigate object recognition and classification in a<br />setting with a large number of classes as well as<br />recognition and identification of individual objects of<br />high similarity. Real-world data sets were obtained for the<br />classification and identification tasks. The considered<br />classification task is the discrimination of modern coins<br />into several hundreds of different classes. Identification<br />is investigated for hand-made ancient coins. Intra-class<br />variance due to wear and abrasion vs. small inter-class<br />variance makes the classification of modern coins<br />challenging. For ancient coins the intra-class variance<br />makes the identification task possible, as the appearance<br />of individual hand-struck coins is unique. Figure 1 shows<br />sample images for the considered collections of coins.
Huber-Mrk, R., Nlle, M., Rubik, M., Hdlmoser, M., Kampel, M., & Zambanini, S. (2012). Automatic Coin Classification and Identification. In Advances in Object Recognition Systems. InTech. https://doi.org/10.5772/35795