TOC | What CHI Does | The Big Picture | Mission | Contact | Bios | Acknowledgements


Empirical Provenance :
A New Paradigm for Imaging Authenticity

What is Empirical Provenance?

The key to widespread adoption of digital representations is simple evaluation of the representation’s trustworthiness.
		Empirical – Information about the world derived from sensory experience
		Provenance – The known history of something in our world
		Empirical provenance – The history of all actions performed on empirically captured information about the world
		Describes the process history pipeline applied along the path from raw data acquisition all the way to the completed representation

Making a Case for Empirical Provenance

Empirical provenance techniques could have helped scholars and revolutionaries in the early 20th century when Russian dictator Josef Stalin was obliterating all traces of his rivals for Communist Party leadership. In this crowd shot example, Stalin erased the image of popular Red Army leader Leon Trotsky (below, left) from his position next to Bolshevik icon Vladimir Lenin (below right).

In the hypothetical example shown here, using empirical provennce techniques would have revealed what had been done to the original photo. If the original photo was available, along with the empirical provenance record, the trustworthiness, or lack thereof, would be clear to all. (see photo-illustration of possible empirical provenance Photoshop Action palette, right).

If present in digital representations of cultural heritage materials and associated with the original acquired data, empirical provenance can:
		Permit the qualitative evaluation of digital information
		Allow digital information accuracy to be tested through confirmatory replication
		Increase the acceptance of online information by scholars, educators, and the public
		Promote collaborative, distributed scholarship
		Reduce risk through transparency
		>Analogy: ingredients labels on food.

Empirical provenance democratizes and decentralizes the empirical capture of our cultural heritage by adding these elements: If joined with digital representations of cultural heritage materials, empirical provenance can:
		Decouple authenticity from authority
		Deliver a transparent process to replace reputation in reliability assessment
		IAllow anyone to create scientifically reliable documentation by following proper practices
		Empower cultural communities to care for and share their heritage.

RTI is an example of a class of robust digital representation methods that are:
		Based on digital photography
		Inexpensive
		IEasy to learn
		Capable of automatic post-processing
		Capable of automatic recording of empirical provenance
		Capable of automatic mapping to semantically robust archiving schema (i.e. ISO standard 21127, the CIDOC CRM).**

Empirical Provenance Examples

Non-digital – Evaluating the trustworthiness of a plaster model replica of an architectural feature requires this data:
		Surface preparation
		Type of flexible mold material
		IType of structural shell
		Final finishing of the mold
		Material composition and casting method of the final model
		Chasing (finishing) of the plaster model (This information permits assessment of surface fidelity and the probability of warping during casting.)

Digital photography – Evaluating the trustworthiness of a digital photograph requires this data:
		EXIF* and IPTC* data
		Raw conversion settings
		IPhotoshop actions and settings
		Distortion correction
		Image resizing
		Sharpening
		More ...

Reflection Transformation Imaging (RTI) -- RTIs built from image sequences:
	Processing is identical for each image
		Data saved:
			Light position determination
			Original raw images
			RAW to TIFF conversion settings (.XMP files)
		All image processing done with CHI’s PTM generator program (fig 1) which saves:
			Photoshop actions and settings (including resizing and sharpening)
			Light position files and command line arguments for RTI generation software
			Log file (fig 2) of every action it performs
			Final RTIs in various sizes

	figure 01 (left), figure 02 (above)
This information, used in association with the original RAW capture, permits assessment of the impact of automatic and volitional image editing operations on the images “authenticity.” Then you can answer questions such as: “Can I use this digital representation in my work?”