piXserve-P: Key Features


Automatic Indexing

Point piXserve to a repository of video files or to live IP addressable video cameras to automatically index their contents.  No manual intervention or data entry required. piXserve “sees” what is in the image and automatically creates records that describe the discernible objects in the image.  This includes faces and text strings in a variety of languages. 

When piXserve detects/recognizes a particular object class, a searchable keyword (a "tag") is automatically added to the set of records that describe the item in that image.  The list of automatically recognized objects is in the thousands, and it continues to grow.   Even if the item is not recognized, users can still search for it based on visual similarity criteria.  If the image contains embedded information (EXIF or IPTC records), that information is also captured and can be searched.

piXserve utilizes all the compute resources available on the server (the CPUs, the cores on each CPU, and those in attached GPU cards).  piXserve-P can also leverage compute resources from other piXserve instances available on the network.  The work (both indexing and searching) can thus be spread across all the nodes that piXserve-P is connected to.  This provides for significant scalability, achieved by simply adding worker nodes to the configuration.

piXserve can process over 90 image formats and most standard video formats.  Typical output formats from today's live video security cameras are also supported.  In the case of live video, the administrator assigns indexing resources to a stream (for exampl, one or two CPU cores), and piXserve indexes all it can based on those compute resources and what the camera is outputting (many cameras output only a few frames a second).   Also, if the video from a live source is being simultaneously recorded by a capture system or video-management system, piXserve-P provides the capability to correlate the real time records it processed to the files being separately saved.  This allows for playback of video search results.


Powerful Search

Through a web browser interface, users login to piXserve, connect to available databases and formulate search queries to retrieve images and specific segments of video files:

  1. Use an arbitrary image from user’s dektop to search for images that contain the same or similar items
  2. Drag and drop a picture of a person, and search for images/video segments where that speficic individual is also present
  3. Use the mouse to point to an area of the query image to indicate which specific item(s) or face(s) should be searched for
  4. Browse the contents of existing databases, and use any of those images to formulate a visual search query
  5. Search images by automatically generated keywords
  6. Type a text string to search pictures where that string appears in the field of view (a license plate, a street name, text on a t-shirt, scrolling text on a TV broadcast, etc.)
  7. Perform not only simple but also complex searches (Example: type a text string of a restaurant name that is visible on the screen, and combine that with a picture of a person in order to find video segments where both are in the same frame. Use AND, OR, and NOT operators to combine up to 6 criteria in a single query
  8. Search by file name (or any text sub-string for that name)
  9. Search by GPS or other external metadata, if available
  10. Use piXserve-suggested keywords to describe the contents of a picture


A desktop and a simplified mobile version of the piXserve user interface are available.  Both run in HTML-5 compatible browsers, so no plugins or apps need to be downloaded.  A REST based Web-Services API is also available.  Customers use that interface to integrate piXserve functionality into their existing workflow applications.  For more information about that, please click here.

Comprehensive Search Environment

Users can prune and save search results for future reference. Links to search results can be saved as XML and exported for ingest by other applications.  Users can interact with the software to improve searches, assign names to recognized individuals, provide examples of specific named objects to keep track of, and many more features designed to help the productivity of your work.