Digital Asset Management – Part 3

Digital recording devices such as digital cameras and camcorders have become available and accessible to the general public.  The word, “file” has taken on a new meaning. Digital image files were being created at an astounding pace.  Industries slowly abandoned older processes for the new digital media.  Standardization became more important than ever.  Record management (including image management) has become a profession.  The need for standardization has been recognized by governments as an issue to be addressed.  New methods of IR needed to be developed.  So how do we manage this information?  A system must be put in place.  For a system to be successful, it first needs to be analyzed.  Questions need to be answered.  Who is the indented audience or end user?  What is the purpose of the system?  How will it be accessed and stored?   “…The primary goal of a records management system is to assist in the effective and efficient management of…records and to provide…the clientele with relevant information at the right time…”(11)

As the field of digital information retrieval expands exponentially, subgroups have emerged.  Image Management is one of those fields.  Hardware and software is being developed daily to keep up with the demand for efficient reliable means to manage digital images.

The algorithms developed to retrieve the vast amounts of digital information fall into two basic categories – the description-based approach and the content based approach.(12)  The description based approach requires human interaction.  Information such as keywords and names are input and stored so that someone may retrieve the information at a later date.  The content based approach involves, “Automatic processing of multimedia information.”(13)

The description-based approach to image classification encompasses a huge amount of information, but is necessary for making the images useful.  Without this, finding an image can become like finding a needle in a haystack.  In 1999 John P. Eakins and Margaret E. Graham published a report which classified image attributes into a three level hierarchy.(14)  Eakins and Graham present a good structure to get the thought process started when developing an image management of your own.  The three level hierarchy is described below.

Level 1: Primitive features (color, texture, shape – e.g. red triangle)

Level 2: Logical features about the identity of objects (general or specific – e.g. office building, Sears Tower)

Level 3: Abstract attributes with high level reasoning about meaning and purpose of object or scene (e.g. “aboutness of images” – festive atmosphere in a party scene)(15)

Levels 1 and 2 can be handled fairly automatically.  For example, to retrieve images by color, the computer would, “look” at images in the database according to the histogram.  Texture could be pulled out in a similar manner by calculating, “The relative brightness of selected pairs of pixels from each image.”(16)  Level 3 requires much more human interaction to function.  Software programs have been developed to address all three levels.  It is vital to have consistency.  There are programs that help standardize categories and data to be input to those categories.

Reblog this post [with Zemanta]

Leave a Reply

Name (required)

Mail (will not be published) (required)



Current ye@r *

WP Flex by WP Queen
Wordpress theme developed by Simpler Computing and others - Wordpress and WPMU Plugins, custom code and more.