This paper shows an interesting snapshot of online visual search engines, analyzing some recent examples: Behold, Tiltomo Beta, Piximilar, and Riya Beta. With some examples, the semantic gap--the gap between the description of the class of images we want to extract and what we actually obtain from the search engine--is highlighted. In particular, the added value of performing text recognition on the images is presented.

What do we get from the paper? In my opinion, it introduces the main problems related to image indexing and retrieval with simple examples. It is also an invitation to learn more, by reading more technical papers. The language is easy to understand and the layout of the paper is enriched with several clarifying illustrations.