Centering Lab · measured, not estimated

How it works

Centering Lab measures opposing printed borders after perspective correction. The engine proposes the physical outline and printed lines; you confirm them in two steps before reading horizontal and vertical ratios for each face. The result includes placement uncertainty and centering ceilings. Missing frames, glare or unclear cuts can require adjustment or prevent a reliable measurement.

Check a card

Card centering API · curl · JavaScript (Node) · Python · OpenAPI

Start with the physical card outline

First, confirm the four proposed corners on the original photo. Centering Lab then maps the quadrilateral onto a rectangle. This perspective correction straightens a tilted photo. The original image supplies the rectified pixels; a smaller copy is used to search for the outline.

A plastic sleeve, toploader or binder pocket can create a second rectangle. Check the cyan lines against the actual cut of the card. Remove rigid plastic before measuring.

Two independent border signals

The color method looks for a sustained change from the outer printed border. A second method looks for a continuous gradient ridge across many scanlines. Agreement supports the measurement; disagreement increases the uncertainty.

Text, foil reflections and artwork crossing the frame can confuse both methods. The lime lines are the best detected starting points. They need manual review when the border is not read precisely.

Measure the most off-center point

Each axis is measured at three stations: 20%, 50% and 80% along the straight edge. For each pair of borders, the wider border divided by their total gives the major percentage. The app keeps the most off-center station.

Left/right and top/bottom stay separate. BGS can require different limits for the two axes. Rotating the image exchanges the axes and preserves their ratios.

What the ± means

After your confirmation, ± estimates line placement precision from the original card resolution and the visible cut transition width. Enlarging a low-resolution image does not create precision. It does not certify that the chosen printed frame is correct.

Every photo, including the sample, follows two steps: confirm the card edge, then check and confirm the eight lines. Grades appear only afterward. Amber edges indicate detector uncertainty; they guide your review. Adjust lines reopens step two and lets you return to the outline.

A ceiling, not a predicted grade

Centering Lab measures centering and shows the best grade that centering allows. A photo of the borders cannot assess corners, edges, surface defects or authenticity. A good centering ceiling does not promise that grade.

Full-art designs may have no usable printed border. Measure the reverse or set a clearly identifiable frame manually. Artwork is not automatically treated as the physical card edge.

Take a useful photo

Place the card flat on a dark, matte surface. Photograph it from straight above, with all four corners visible. Use diffuse light and remove the sleeve or toploader when possible.

For photos with a background, fewer than 600 pixels across the card may prompt a closer shot. Cropped scans use the whole image and do not show toploader or distance warnings. Keep the image sharp and avoid digital zoom, glare and strong perspective.

The photo never leaves your browser. No account is required. The analysis runs locally in a worker; exported results are created on your device.

For a practical walkthrough: How to measure card centering with a phone or ruler · Is 60/40 centering good? Card centering ratios explained · Front vs back centering: which side limits your card?.

Sources

Editorial policy and corrections

Check a card