In order to explain this strategy we will firstly explain an important new concept called Basic Light Responses (BLR) and how Master Stones could be used to grade each of these BLR factors. With this understanding it is possible to show how an accurate 3D model of the diamond under consideration can be used by software systems to grade its cut quality.

a) Basic Light Responses (BLR)

Examples of Basic Light Responses (BLR):

• Light return to the observer's eye;
• Light leakage though the pavilion;
• "Nail head";
• ETAS - part of space visible trough moving diamond

A value for each BLR can be computed on the basis of a 3D mode.

The value is individual for each and every diamond and characterizes its cut.

Two types of such responses should be considered:

1. Positive responses: e.g. scintillation, brilliance, fire, and contrast
.
2. Negative responses: e.g. "fish-eye", "nail head", light leakage
.

b) Master stones for cut grading

The purpose of the Master Stones Set (MSS) is to:

• Demonstrate appearance differences between diamonds of different grades
• The "good diamond" or "bad diamond" concept will be applied to each positive and negative effect
• There is no objective definition or formula for 'beauty'. It is necessary to accumulate and to analyze human perception data for each BLR.
• Each Master Stone will be graded by different experts such as manufacturers, buyers and sellers. The experts will be chosen for their sensitivity to nuances of diamond appearance.
• Human experience in judging diamonds and designing new cut styles is geographically dispersed. It is unlikely one organization can prepare this Master Stone set. The participation of various organizations to approve the set will prevent foolish errors.
• Involvement of a complete range of manufacturing specialists and cut designers is most important when making such a set.
• Manufacturers interested in developing, inventing and promoting new cut styles have the resources to add many new fancy shape stones to the MSS each year.
• Properly graded new cuts will quickly populate new market niches. Naming protection can be provided by agreement from Labs, without the expense of patents and trademarks. Manufacturer's innovation and value adding will prevent diamonds becoming a commodity.
• The Master Stone Set will include a variety of diamonds: good and bad, traditional and non-standard, round and fancy etc. As desirable and non-desirable effects become better understood, more innovation will lead to new cut styles and 'art proportions'.
• The MSS will be used to verify or ratify ANY or ALL cut grading systems; all systems should give the same grades for stones that represent examples of specific good or bad BLR responses e.g. very good scintillation and very poor scintillation.
• The organization that holds the MSS will make them available for institutions and laboratories involved in cut studies and grading.
• Verification and validation of new grading systems will lead to greater consumer confidence and avoid contradictory grades.
• New cuts may appear that lead to conflict of opinions of experts' or various grading systems. If grade variations are found to exist, then one or more systems may need revising.

c) Cluster approach to BLR data

The Master Stones Set will be divided into different BLR groups based on expert human grades; e.g. brilliance, scintillation, fire, and contrast. The same process will be performed by the computer based 3D model cut grades.

As a result we receive two sets of data that can then be compared: (1) Computation-based clusters and (2) Expertise-based clusters (fig. 3).
A proposed method for correlation between these two cluster maps is via 'Neural Networks', a special statistical method for comparing data arrays to build a common system.

The BLR-based grades and the experts' grades may match or not match:

1. If they do match, then the system is complete.
2. Clusters in the BLR grading match more than one cluster in the expert human observations; the match is not adequate.
3. Clusters have inadequate matching.

In this way the Master Stone Set may be used to verify various cut grading systems and 3D models and the ETAS approach may be used for both Cut Grading and devising new cuts.

This approach is robust against errors in both diamond expertise and BLR computation.

BLR Calculation		Diamonds separation by human observation

Tolkowsky	Hight LightReturn	NailHead
Good	Brilliance	Kozibe
Bad	Fire	FishEye

	Direct correlation
	BLR set is enough for cut grading
	BLR set is not enough for cut grading
Figure 3. Clasterization