Description
In the project, we developed and characterized a digital image processing algorithm for the automated detection of beer labels from photographs of 100 different beer bottles. The algorithm achieved a high (100%) success rate, was sensitive to subtle differences between distinct labels, and displayed robust classification against simulated camera motion and large camera-to-bottle distances. This tool would be appropriate for various mobile phone applications, including resources for consumer product information and even social networks.
Image processing strategy for SIFT-based beer label classification
Reviews
There are no reviews yet.