Picket Fence Analyzer AI

Automated MLC leaf analysis with MLC type and treatment mode detection (static / VMAT) and ANSM compliance.

Picket Fence images (DICOM)

One or several EPID images (up to 12), e.g. one per gantry angle.

MLC type

QC periodicity (ANSM)

Tolerance (mm)

ANSM value: 0.6 mm.

Action tolerance (mm)

Capped at the tolerance (pylinac rule).

Nominal gap (mm)

Gap width programmed in the QA plan.

Detection settings

Number of pickets

Leave empty for automatic detection.

Orientation

Invert the image

Crop the edges (mm)

Median filter size (px)

Sag adjustment (mm)

Leaf analysis width ratio

FWXM (%)

Height threshold

Edge threshold

Required prominence

Report

Author (report)

Comment (report)

Engine: pylinac PicketFence. Your files are deleted within one hour. Do not upload identifiable patient data.

Analysing the images…

No analysis yet.

Upload one or several Picket Fence images, or run the demo image, to see the results here.

How it works

Upload one or several Picket Fence EPID images — ideally one per cardinal gantry angle for the ANSM monthly control.
The MLC type and the test mode (static, dynamic, VMAT) are read from the DICOM header; pylinac finds every picket and measures the position error of each leaf pair.
Each file gets two statuses: against your tolerance, and against the ANSM criteria of the decision of 28 February 2023 (max error ≤ 0.6 mm and at least 95 % of leaves passing).
Optionally, an AI expert drafts an ANSM-oriented interpretation from the measured values — always to be validated by a medical physicist.