Intraoral scanning of the palate is considered reliable for human identification;
however, its accuracy on postmortem tissue remains dubious. This study aimed to investigate
the effect of tissue decomposition on the precision of the intraoral scanner and the
deviation of the scan. Ten fresh lamb (Ovies aries) maxillae were either unwashed
or washed, selected, and stored at 20.5 °C and 80% humidity for 20 days. Each palate
was scanned three times a day with an Emerald S intraoral scanner. The anterior rugae
area was cropped for analysis. The three scans of each day for each lamb were digitally
aligned using the iterative closest point algorithm to ensure precision. The day one
mesh was compared to each subsequent day to assess the postmortem scan deterioration,
and a quadratic curve was fitted to the data. The mesh from different lambs was compared
on day one to calculate the differences between the lambs. The length, location, and
value of the largest curvatures of five randomly chosen rugae on each specimen were
determined. A supervised machine learning procedure using linear discriminant classification
assessed the specificity and sensitivity of singular ruga discrimination. Precision
was significantly lower (p < 0.001) in the unwashed group (0.025mm) compared to the
washed group (0.013mm), but the postmortem days had no effect. The deviation curve
for the unwashed samples had a significantly higher quadratic term (p < 0.05) compared
to the washed sample, indicating a slightly greater deterioration after day 11. The
least difference between lambs was 0.484mm. The deterioration curves crossed the minimum
value on day 6 in both groups. The sensitivity of rugae detection was 0.89 on day
one and decreased to 0.69 on day 20; the specificity ranged from 0.59 to 0.66. Intraoral
scanning is an accurate approach for postmortem palatal imaging. Superimposition of
the anterior palatal scan can accurately distinguish between lambs for up to six days.
Nevertheless, deteriorated rugae can still be distinguished with moderate accuracy
for up to 20 days.
