Frequently Asked Questions

This is almost always caused by scale calibration, not a segmentation error. Because area is a squared value, even a small error in the plan's scale gets amplified significantly in the final number. Don't rely solely on auto-detected scale — use the Manual Calibrate tool: find a known dimension on the plan (for example, a wall labeled 10'-0"), click its start and end points, and enter the real-world distance. The system recalculates the pixel-to-foot ratio from that input.

This usually means the crop is too wide (including too much whitespace) or the plan's contrast is too low for the system to read the wall lines clearly. Try tightening the crop in the Plan Prep stage to exclude margins, and check whether the source PDF is a faint scan — increasing contrast in a PDF editor before uploading often helps. If rooms still aren't detected, use the Draw Room tool to manually trace one room; this gives the system a scale "hint" that can help it auto-complete the rest of the plan.

A room polygon with no finish material assigned is treated as null data, and the system skips unassigned shapes to avoid cluttering the export with empty entries. Before exporting, make sure every active room has a finish code assigned in the editor. Use the Validation Check button, if available, to highlight any unassigned polygons before you export.

It's recommended for complex files. Advanced PDFs often contain multiple layers (transparency, annotations, etc.), and flattening merges these into a single image layer, which is typically easier for the segmentation engine to process without errors.

Yes. A plan that's rotated even slightly — a couple of degrees is enough — can cause a "stair-stepping" effect on wall lines (a form of aliasing), which confuses the Solidification engine responsible for closing broken lines into clean shapes. If your scanner fed the paper in at an angle, use a deskew tool on the file before uploading.

A vector PDF is generated directly from CAD software and contains real mathematical lines and text, which the system can read with the most precision. A raster PDF is essentially a photograph of a paper plan — pixels only, no underlying data — and relies on OCR and Solidification to recover room geometry. Both are accepted, but a vector PDF, where available, will generally produce more accurate results with less manual correction.

Scale calibration is the process of verifying the pixel-to-foot ratio — the constant that defines how many screen pixels represent one real-world foot — using a known dimension on the plan. Because every area calculation depends on this ratio, and area is a squared value, even a small calibration error compounds into a significant error in the final square footage.

Solidification is a pre-processing step that runs before segmentation to fix broken or faint lines in a scanned plan — connecting gaps and filling them in so the system can recognize a room as a fully closed shape, rather than misreading a broken wall line as an opening to the outside.

Circuit Flow is Estoro's proprietary five-step linear workflow for processing a floor plan: Ingest, Select, Scale, Crop, and Segment. It's the underlying pipeline behind every takeoff generated in the platform.

A waste factor is a percentage added to a room's net calculated area to account for material lost to installation cuts, breakage, and mistakes during the actual flooring job. For example, a 100 sq ft room with a 10% waste factor would require 110 sq ft of material to order.

A polygon is simply a closed shape with multiple sides — the basic geometric unit used to represent any room's floor area, in either mode. Boundary mode is a specific editing view where that polygon is shown as a thin outline (its boundary) with a transparent center, rather than filled in solid, which makes it easier to work with open-plan spaces and material transitions.

OCR (Optical Character Recognition) is the technology that converts pixels of text in a scanned or image-based PDF into actual, readable text characters. In the platform, it's used specifically during the legend-extraction step, to read finish schedule tables and symbol codes from the plan.