VePAIR (Visual Engines Photo Archive Information Retrieval) is a software solution for the management of large (millions) photographic archives. The VePAIR core technologies are Visual Information Retrieval and Large Scale Image Indexing developed by the Networked Multimedia Information System (NeMIS) laboratory of ISTI-CNR in Pisa.
- Search for similarity
- Search for duplicates
- Brand recognition
- Face recognition
- Duplicates removal
- Grouping by similarities
- Grouping by content
- Filtering of newly acquired photos
- Copyright infringement verification
how it works
SEARCH BY SIMILARITIES
This is a very important features in the VePAIR solution. It allows to identify groups of images containing similar objects such as, for example, the same monument, the same painting, etc. It is important to note that object matched during a similarity search doesn’t need to be identical; PIAR matching algorithm allows to recognize images even when partially distorted, with a different dimension or with a different daylight exposure.
SEARCH A COPY OF AN IMAGE
Searching a copy of an image using VePAIR means searching exactly a specific image with tolerance on the context in which it appear (for example inside a newspaper page), the dimensions (i.e. proportionally smaller or bigger) and color scheme (i.e. transformed in b/w). The copy detection service of VePAIR is the basement of the Copyright Infringement Detection.