You are operating an online shopping website. A user comes in to your site
to search some products by keyword searching or browsing. He or she just finds
something interesting. It could be a bag, a pair of shoes, or a beautiful watch.
The user may want to see more items visually similar to this one. Here
our product, vSimilaritySearch, comes in to play. Check out our
visual similarity search demo.
3. How it works
Given a large numbers of images, we analyze and extract all finger-prints of those images such as color distribution, shape characteristics, and image features. Those images finger-prints are then saved into a special index and database. During search time, an input image is analyzed to extract its finger-print and then compared with fingerprints of image index and database, thus find visually similar images. Thanks to our advanced search ranking algorithms and highly parallel processing, search results can be delivered in a fraction of seconds from ten to a hundred millions of images.