View your test with adjustable white light to easily see where stones are located.

Sherlock Holmes 4.0, Dr. Watson, Sherlock AI & Watson AI

After testing, or after saving a test, open the RESULT screen. You can now add adjustable white light to easily see where the stones are located and control the amount of white light.

On the bottom right of the RESULT screen, you can see the 2-screen icon .

Tap on it and you will get a slider. Use the slider to add white light to the picture.

In the Gallery, where you can see saved tests, go to any saved test, open the RESULT screen, and the slider will be at the bottom of the picture.

Please watch a short video here.

Any test that you wish to make a report for must first be saved.

Then, go to the Library within the App, or on your computer, go to your back office at cloud.yehuda.com and click on the test for which you wish to prepare a report.

Perform the AI analysis. Once the AI is done, click the printer icon or the PRINT button, and choose PRINT or SAVE.

You can choose to get a report with or without the AI results.

This is a great tool for you and for your customers.

You can choose to print or save the report with or without the AI results.

• Multi-Stone Detection
  • analyzes multiple loose or mounted stones.
• Small Stone Identification
  • detects very small stones.
• Manual Marking & Re-Analysis
  • manually mark a stone and run the AI again.
• Per-Stone Confidence
  • view the AI confidence level for each stone.
• Adjustable Confidence Level
  • change the confidence threshold in the cloud.
• Why Sometimes Does the AI Not Recognize a Stone?

The AI is an advisor, not the final authority.

When you and the AI agree — perfect.

When you disagree — investigate the stone and make your final conclusion.

Please feel free to contact us for advice.

The AI model has two stages:

1. Recognize what is a stone.

2. Identify what type of stone it is.

The accuracy of the second stage is above 99% under proper testing conditions.

1. Recognize What Is a Stone

The new AI model can automatically recognize loose stones as small as 0.01 ct. in Watson AI and Sherlock AI.

For mounted stones, it can recognize approximately:

• Watson AI: 0.01 ct. and up.
• Sherlock AI: approximately 0.02–0.03 ct. and up.

If a stone is not automatically found, you can always manually mark the stone and analyze it — regardless of its size.

Why Sometimes Does the AI Not Recognize a Stone?

The most common reasons are:

• The stone is out of focus – Retest with focus.
• Side stones are not facing up – reposition them, so they face the camera, and retest. Make sure the focus is correct.
• Stones are touching or piled up – separate them and retest.
• Pavé settings – stones that are very close together may not be recognized.

Any stone that is not recognized can be manually marked and analyzed.

Please see “Manual Marking of a Stone” below.

2. Identify What Type of Stone It Is

In our internal testing of over 100,000 stones, the AI model achieved better than 99% accuracy.

Real-life results may differ.

Why Sometimes Does the AI Give the Wrong Classification?

There are a few situations that can adversely affect the AI reading.

In most cases, the user can easily identify these situations and understand why the AI may have produced an incorrect result.

Reflections

The light sources are located near the camera. Sometimes a stone reflects the light instead of producing fluorescence or phosphorescence.

The AI may interpret this reflection as the stone shining.

This can occur on any testing screen: COLORED, GLOWING, and SIMULANTS.

Solution: Move the loose stone or jewelry to another position and test again.

The “Flashlight Effect”

Some stones shine extremely strongly — almost like a flashlight — and can illuminate the stones next to them.

This is most common on the GLOWING and SIMULANTS screens, but it can also occur on the COLORED screen.

The AI may then think that a nearby stone is shining when it is actually being illuminated by the neighboring stone.

Loose stones: Remove the very strongly shining stone and retest.

Jewelry: The stone usually cannot be removed, so the result should be evaluated using human judgment.

When reflections and the flashlight effect are eliminated, AI accuracy is significantly higher than 99%.

3. Natural Diamond vs. Moissanite

There is one important area where extra attention is required.

Some small natural diamonds, when mounted, may not show fluorescence. They can therefore behave similarly to Moissanite and may receive a Moissanite result by the AI.

This usually occurs in jewelry containing many small diamonds, where most stones show the expected response but a few appear different because they are not showing fluorescence.

Recommended Solution

Check questionable stones with a loupe and/or a simple diamond pen tester to confirm whether they are diamonds or Moissanites.

Remember

The AI is a powerful advisor — but when a result does not make sense, investigate the stone and use your professional judgment.

Every AI marked stone has its AI confidence level.

• 99% = very high confidence
• 80% = high confidence
• Lower percentages = less certainty

The default threshold is 80%. Stones below 80% appear in black, but their confidence score can still be viewed.

The confidence threshold can be changed in the cloud after saving the test.

🔵 Blue – Natural Diamond

🔴 Red – HPHT

🟢 Green – CVD

🩷 Pink – CZ / Other Simulants

🟠 Orange – Moissanite

⚫ Black – Refer / AI uncertain; move and retest

Change the confidence level of the AI results

After saving a test, on your computer go to cloud.yehuda.com and log in using your app credentials.

1. In Gallery, select a test and click Generate AI.
2. Once the AI results are generated, click/tap any stone to view its AI certainty (confidence) score.

How to interpret the score

• 99% = maximum confidence
• 80% = still a very high-quality result
• Lower percentages indicate less certainty and should be interpreted with caution.

Default filtering

• By default, results are filtered at an 80% confidence level.
• Stones that receive below 80% are marked in black.
• You can still click a black-marked stone to see its confidence score.

Changing the confidence level

• You can change the confidence threshold by adjusting the confidence slider.
• Example: If a stone’s certainty is 70%, it will be marked in black at the default 80%. If you lower the slider to 70%, the AI v result will be shown normally.

Please note: Currently, after generating AI on your phone, you can tap any stone and see its confidence score. Changing the confidence level (threshold) is only available after saving the test and using cloud.yehuda.com on your computer.

False Positives
The AI may occasionally mark metal areas or reflections.
These markings should be disregarded.

Important notes

• Detector vs. AI accuracy:
  • Yehuda detectors will not miss a lab-grown diamond when the final decision is verified by a human.
  • A small percentage (about 2%) of natural diamonds are detected as LGD (false positives) by the detector. In those cases, the AI model will also classify them as LGD, because it relies on the detector’s physical reading.
• AI accuracy:
  • Currently about 99% in our internal tests, and the model is still learning and improving.
• Note:
  • The AI model serves as an advisory tool, not a final authority.
  • Continuous updates and training will further enhance its performance.