Gish vsCalc iPhone/iPad App Reviewed

Published Mar 17, 2011 00:00 AM
PUBLIC - This article does not require an IPVM subscription. Feel free to share.

In this note, we examine a new iPhone/iPad application from Gish Technologies for calculating video surveillance. The goal of the application is to help designers to select the optimal cameras and lens combination for a given scene's requirements. We start with a review of how it works and conclude with a 12 minute video screencast sharing our experiences using the application.

The applications are based on 2 key elements:

  • 4 pre-defined images/scenes: a face, license plate, car and truck: These images demonstrate potential video quality. Users cannot upload their own, use iPhone/iPad2 integrated cameras, etc.
  • Detailed technical information on vendor's camera lineups: Each version of the application comes prepopulated with a different vendor's camera information, currently Axis [link no longer available], Arecont [link no longer available] and Sony

A designer chooses a potential camera, the scene desired (e.g., seeing a face) and the level of visual clarity that is desired. The application then returns how wide a Field of View can meet these specifications. Additionally, the user can enter how far the target (e.g., face) is from the camera. The application then generates the correct lens size needed.

Visual Acuity Index

Perhaps the most interesting and potentially challenging aspect of the tool is its introduction of an new eponomously named index for surveillance video quality [link no longer available]. Rather than use pixels per foot (e.g., 48 pixels per foot for clear faces or 10 pixels per foot for detecting a car), the application uses a scale of 0 to 1 to reference video quality. Inside the application, users can choose between .43, .53, .63 and .73 Gish levels roughly mapping, respectively to detection, recognition and identification levels. We found that Gish levels map to pixels per foot and that the pixels per foot doubles for each level up (e.g., if .43 Gish ~ 10 pixels per foot for a given object type, .53 Gish ~ 20 pixels per foot for the same object).

Using a singular quality scale has precedent in other industries (e.g., Mean Opinion Scores in the audio / telephony markets). That noted, it will interesting to see if this catches on in the surveillance world.

Below is our video screencast demonstrating key features of the applications:

For an alternative to this application, you could use Theia's lens calculator [link no longer available] that includes pixels per foot calculations and sample quality images (see our related lens calculator training report). While the Theia app provides greater flexibility, it also demands greater expertise and time to look up camera specific details.