Skip to main content
The cross-reference system compares a case against records of the opposite type (for example, a missing person against unidentified remains) and returns ranked research leads. Every lead links back to its original record.
Ranked suggestions are research leads, not identifications. Verify any lead with the original agency or source.

​
How leads are scored

Each lead has a score from 0 to 100, plus details that explain it:
  • Confidence: strong, moderate, or weak. Confidence is limited by evidence strength and lowered by discrepancies, so demographic overlap alone is never strong.
  • Evidence score: how much physical and forensic detail overlaps, such as marks, belongings, and identifiers.
  • Breakdown: points per factor, showing whether each one is a match, partial, none, or unknown (when a record lacks the data).
  • Coverage: the percentage of factors both records had data for.
  • Discrepancies: details that point away from a match, such as a large age or height gap.

​
Tune scoring

Signed-in members can adjust scoring weights under Tune scoring to emphasize what matters to them. Tuned results show both your score and the default score.

​
Compare side by side

Place 2 to 4 records next to each other with differences highlighted, so unknowns are as visible as overlaps. Use AI case analysis to explain why two records were linked. All AI output is an unverified research aid.

​
Experimental photo research

Optionally compare an authorized target photo with a selected public case photo to surface a research lead. Paid plans also include experimental age estimation.
Similarity scores are not probabilities, identifications, or proof of a match. CaseDock does not generate age-progressed images.