Currently only Pokémon cards can be verified — other card games are not supported yet.
Computer Vision & Physics

The Science of Card Metrology

Bite-sized engineering deep dives into projective geometry, subpixel signal processing, and how we measure a card from a photo.

Projective Geometry

Why Phone Photos Lie About Centering (And How Homography Fixes It)

How a 5° hand tilt distorts your card's border ratio by up to 8% — and the projective geometry that straightens it out.

“Under a projective transformation, parallel lines converge and relative distances between collinear points are distorted. To recover true Euclidean ratios, the plane projective distortion must be rectified.”
— Richard Hartley & Andrew Zisserman, Multiple View Geometry in Computer Vision (2nd ed.), Cambridge University Press, 2004
Read article →
Surface Analysis

Hunting Scratches on Holo Foil: The Magic of Steerable Filters

Why standard computer vision sees thousands of fake scratches in card glitter — and how directional filter banks extract real defects.

“Directional derivatives of any orientation can be synthesized by a linear combination of a small basis set. Coupled with multiscale Hessian eigenvalue analysis, this separates elongated curvilinear features from isotropic background textures.”
— William T. Freeman, Edward H. Adelson & Alejandro F. Frangi, IEEE TPAMI (1991) & MICCAI Medical Image Computing (1998)
Read article →
Edge Metrology

The 0.05 mm Difference: Why Subpixel Math Decides a PSA 10

How parabolic interpolation on Gaussian edge profiles turns discrete camera pixels into micron-level metrology.

“By computing continuous orthogonal profile derivatives and 3-point parabolic interpolation, true physical edge boundaries can be localized with sub-pixel and sub-millimeter precision.”
— Frédéric Devernay & Carsten Steger, Sub-Pixel Edge Detection & Curvilinear Metrology, INRIA / IEEE TPAMI, 1995 / 1998
Read article →
Modern Geometry

Beyond RANSAC: How MAGSAC++ (CVPR 2020) Solved Card Corner Jitter

Why 1980s consensus algorithms produce wobbly card boundaries, and how modern marginalization achieves rock-solid subpixel alignment.

“By marginalizing over the noise scale instead of applying an arbitrary hard inlier threshold, MAGSAC++ achieves geometrically optimal parameter estimates with high speed and zero manual tuning.”
— Dániel Barath, Jana Noskova, Maksym Ivashechkin, Jiří Matas, CVPR 2020 (IEEE Conference on Computer Vision and Pattern Recognition)
Read article →
Surface Metrology

The Specular Glare Dilemma: Separating Flash Reflections from Real Dents

How dichromatic reflection decomposition stops bright lighting hotspots from masquerading as card damage.

“Separating specular and diffuse reflectance components allows invariant inspection of surface micro-topography independent of ambient light source direction and intensity.”
— Artusi, Banterle, Chetverikov et al., Computer Graphics Forum & IEEE Transactions on Pattern Analysis and Machine Intelligence
Read article →
Cross-Grading Science

PSA vs. CGC vs. BGS: The Metrology Behind Cross-Grading and Company Strictness

Why a PSA 10 can fail at Beckett, why CGC 9s cross to PSA 10s, and what 14,900+ slabs reveal about centering and surface thresholds.

“Inter-rater agreement on ordered quality categories must penalize large disparities quadratically while isolating systematic company-specific acceptance surfaces.”
— Jacob Cohen & Peter J. Rousseeuw, Weighted Kappa & Robust Metrology in Industrial Inspection, 1968 / 1984
Read article →

Put Computer Vision to the Test

Upload your own card scans or phone photos to receive instant subpixel centering ratios and grade estimates.

Grade a card now — Free Open Beta