Ghost · alpha · worked example
How to compare front and back garment photographs
A garment study needs consistent views before a 3D approximation is useful. Mismatched pose, distance and cropping can make the same garment look like two different shapes.
Updated October 10, 2026 · by thecreatingco
A small example to work through
Our fictional jacket brief has a front photograph 1800 pixels high and a back photograph 1500 pixels high. Compare the shoulder-to-hem span rather than forcing the entire images to match.
| Item | Example | Notes |
|---|---|---|
| Front shoulder-to-hem | 900 px | Reference span |
| Back shoulder-to-hem | 750 px | Scale comparison view by 1.2 |
| Sleeve opening | Partly hidden | Mark uncertainty |
| Inside lining | Absent | Do not infer material construction |
Download the editable text worksheet · Download the CSV worksheet
Use the blank observation and decision fields to record your own result. These downloads are planning worksheets; they do not load an app project or demonstrate a completed export.
Try it and inspect the result
- Choose clear front and back photographs with permission to use them. Avoid strongly different poses and wide-angle views.
- In Ghost compare the visible silhouette from equivalent directions. Keep source photographs beside the approximate 3D study.
- Record which features are visible, obscured or absent. Review a saved result before closing the current workspace; keep the original images independently.
When the result differs
- Front and back lengths disagree
- Compare an actual garment landmark before rescaling the whole photograph.
- One sleeve appears wider
- Pose and perspective may explain the difference; do not immediately remodel it.
- Want a sewing pattern
- Photographs do not recover measured pattern pieces, seam allowance or wearable fit.
What this example establishes
The 1.2 ratio aligns two comparison spans in this invented example. It is not a garment size measurement. Ghost produces approximate visual studies; hidden construction and exact fit need a different source of evidence.
This page supplies a worked example and a review method. It does not report a newly completed end-to-end app test. For current availability and edition details, read the Ghost overview.