Government agencies, law enforcement and private enterprise are all utilizing a new generation of digital imaging and video capture technologies to conduct security and surveillance tasks. Their greatest cost and limitation in taking fuller advantage of the wealth of captured images is the manpower required to decipher what is significant or of interest out of terabytes of images.

The use of visual search enhances the value of the captured media. The piXlogic technology can greatly reduce the number of hours spent reviewing useless content and concentrate resources to review "suspicious" activity.

The technology can be used on both pre-recorded images and real-time video feeds. Some possible uses of the piXlogic software in security and surveillance markets are:

• Real-time analysis of video feeds:
  Video security/surveillance systems (CCTV) can benefit from the application of piXlogic image analysis software. The piXlogic software can be used to carry out a "continuous" search of a live video feed. An operator could formulate one or more standing queries, and the live video can be analyzed in real-time. When the query item(s) are detected in the field of view of the camera, triggers can be activated (send an e-mail, start recording, sound an alarm, etc.).

  For example, a camera monitors the entrance to a building with heavy pedestrian traffic. The software sends an alert each time a person enters or leaves the building carrying an unusually large object (eg, large or long box).

  In another example, the feed from an altitude camera triggers an event when two or more vehicles of a certain type come to an intersection within a predefined time interval. The piXlogic software effectively handles this situation even when there is a lot of activity in the video. For instance, even if there were a tremendous amount of pedestrian traffic, the software would not trigger an event unless it identifies an object that physically matches the search criteria.

• Postprocessing of images/video to find objects of interest:
  Scene analysis is the most common way to segment a video. With today's existing techniques this segmentation is done at the time of ingestion, and is static (i.e., only done once). With piXlogic, video is segmented at the time the user carries out the search with the segmentation based on finding objects of interest.
  
  One possible area this could be applied to is in analyzing air reconnaissance images/video. With dynamic segmentation, it would be very easy to go back and analyze historical footage to find objects that might have recently become more relevant, such as a certain type of tank, artillery, vehicle, or a person that has distinctive characteristics.

• Mine existing image assets quickly and efficiently:
  An existing library of assets (images and video files) can be automatically indexed by piXlogic. The speed of indexing can be scaled to faster than realtime with the appropriate computing resources. The automatic segmentation process can result in an index that captures the logical contents of the images. The index can be searched using images that are either internal or external to the system (e.g. the images that are used to formulate the search query need not be part of the digital asset database. An example of using external images would be a snapshot taken in the field only for the purposes of making a visual query (e.g. the user is asking, "What is this?") and not intended for storage in the image database.