In order to create a
searchable index of the contents of a picture/video
- automatically segments the contents of the image in a way
that much resembles how a person would do it.
- creates a mathematical
description of these segmented
objects "on the fly", and stores
these descriptions as a searcheable index.
- reasons about what it "saw"
in the image and develops an
initial level of "understanding" about the
contents and context of the image. In
some cases it is able to "tag" the items with a descriptor (a
"Notion"). An evolving and increasing vocabulary of
Notions is available in piXserve (ex: sky, vegetation, flower,
face, building, car, map, airplane, etc.)
- uses calculated descriptions, Notions,
relationships and context information (foreground, background,
etc.) to make comparisons
between a query image and images/video frames previously
indexed and finds results that most closely match what the
user was searching for.
Additional key features
of the piXlogic technology are:
- ability to "see" not only visual objects
but also text strings on the image (language independent)
- ability to carry out complex and
multi-modal search criteria (ex: find an image/video segment that
contains something that looks like this item in image 1,
AND/OR/NOT this item from image 2, AND/OR/NOT this specific text
string anywhere in the image).
- ability to create alert criteria that
notify users when content of interest to them is indexed by the
These features and characteristics, and
more, are available in several piXlogic products:
piXserve - A client/server product line rich in
functionality and suitable for workgroups and for demanding
- piXserve-Web Edition
- piXserve-Standard Edition
- piXserve-Professional Edition
- piXserve-Enterprise Edition
- piXserve-Elastic Cloud Standard
An SDK (software developer
kit) which can be used to integrate
visual search functionality into
existing commercial applications. The API is well-defined,
comprehensive and based on industry standard interfaces.