Automated grading uses a repeatable pipeline designed to eliminate subjective human interpretation.
1. Imaging
The card is scanned using:
ultra-high-resolution cameras
controlled lighting
multi-angle capture
laser measurement (in AGS machines)
2. Pre-Processing
The system corrects:
perspective
skew
lighting
noise
3. Feature Extraction
AI identifies:
centering ratios
edges
corners
surface textures
defects
print lines
whitening
nicks & abrasions
4. Defect Classification
Each flaw is measured in:
size
depth
location
severity
5. Scoring
The system combines:
measurements
historical data
defect weights
proprietary grading logic
6. Output
You receive:
final grade
subgrades
heatmaps
defect list
scan report
population integration
This entire pipeline is consistent, auditable, and can be replayed, unlike human grading.