Furthermore, we provide corresponsing evaluation scripts. This dataset is maintained by the Multimedia Computing and Computer What is meant by multimedia Lab, Augsburg University. RootSIFT descriptors and visual words now available for download.
2013: Added section Evaluation Protocol, i. 2012: Added evaluation scripts and sample data. 2012: Added errata for the paper clarifying a minor issue. We collected logos of 32 different logo brands by downloading them from Flickr. All logos have an approximately planar surface. The retrieved images were inspected manually to ensure that the specific logo is actually shown.
The whole dataset is split into three disjoint subsets P1, P2, and P3, each containing images of all 32 classes. To facilitate the development of high-precision classifiers the evaluation of their sensitivity on non-logo images is important. Therefore both partitions P2, and P3 include another 3000 images downloaded from Flickr with the queries “building”, “nature”, “people” and “friends”. These images are the negative images and complete our dataset. A brief summary of the data subsets is shown in the table below.
The binary masks are provided as . Masks are single-channel PNG images of the same size as the original image. Also thumbnail images are included for easy visualization of results. Evaluates the classification of logo images by computing precision, recall and more. Download the evaluation kit including all scripts plus additional sample data. If you encounter problems, please contact us.