2007-11-05

The Open Source Rationale for the I* Metric

An article was recently published in IS&T's Technical Program and Proceedings:  NIP 23:  The International Conference on Digital Printing Technologies entitled, "Evaluating the Image Permanence of Full Tonal Scale Human Skintone Colors in Photographs Using the CIELAB Colorimetry Based WIR i-Star 'Retained Image Appearance' Metric" by Henry Wilhelm and Dmitriy Shklyarov.  This paper is well worth reading and contains important observations about image permanence testing complexities related to modern multi-colorant digital printing technologies.  It is available in PDF format at Wilhelm-Research.com and in the conference proceedings which can be purchased from the Society for Imaging Science and Technology (IS&T).  Unfortunately, the title of the paper refers to the I* metric (pronounced "i-star") as the "WIR i-Star" metric.  This description is regrettable because it implies that the I* metric is proprietary to Wilhelm Imaging Research, Inc.  As the inventor of the I* metric, a developer of test methods for its use, and as the primary author of the original paper about the I* metric published in 2004 (available in PDF format here), I wish to clear up any misconceptions about the proprietary versus open source nature of the I* metric.

The I* metric is a non-proprietary, open source color and tonal accuracy model, freely available for other researchers to use.  It is relatively straight-forward though perhaps not trivial to program into a spreadsheet application like Excel.  Just as it would be inappropriate for me to refer to an industry standard color model like CIELAB, for example, as the "Aardenburg Imaging CIELAB metric", so too, is the use of any company branding with respect to the I* metric.  That said, the apparent confusion in the title of the paper cited above may have arisen because Wilhelm Imaging Research has also been working on a robust software application to exploit the virtues of the I* metric and method.  Adding brand recognition to a proprietary software application is appropriate, but any implicit or explicit description suggesting that the metric itself may be proprietary is incorrect.  In the interest of full disclosure, I am supposed to receive royalties from future sales of the WIR software implementation of the I* metric should it be sold to other companies or to the general public.  However, the non-proprietary integrity of the I* metric itself needs to be respected and nurtured by all concerned.

At this point in time, the I* metric is still too new to have received much exposure in the industry.  It has not yet been vetted or "blessed" to any significant degree by other experts in the field of color science or the imaging science community.  The greatest possible chance for this validation process to occur is when A), the metric and methods are able to be freely examined and implemented by other experts in the field and B), the metric itself is fully understood to be an open source metric that will always remain free for anyone to use.

It took me more time than I'd care to admit to derive the I* metric, and this research was equally sponsored by both Wilhelm Imaging Research, Inc, and my own company, McCormick-Goodhart, Inc.  I enjoyed a great collaborative environment with colleagues at Wilhelm Imaging Research who helped me refine many of the technical details.  I believe in the I* metric.  My new company, Aardenburg Imaging & Archives (parent company, McCormick-Goodhart, Inc), will be utilizing the I* metric to describe how well modern digital printing processes age over time.  I also hope to apply it to other image quality characterizations as well, and I welcome others to do the same.  Yet I am well aware that any new approach to a long standing industry problem simply takes time to gain acceptance.  The eventual acceptance, of course, depends upon many factors, but fundamentally it depends upon whether the new approach truly has merit.  We will find the answer to that question more quickly if researchers who wish to try it are not confused about their right to implement it.