LCD Online

This tool is designed to process DICOM images of the LCD phantom exclusively, and must not be used with clinical images containing patient information (name, ID, birth date, accession number, etc.). 

All image decoding and ROI computation run entirely client-side, in JavaScript executed by your browser — no pixel data, file content, or DICOM metadata is transmitted over the network or stored on any server. However, this does not extend to what you do with the results afterwards: an exported Excel file or a screenshot can still carry identifying information if present in the source image. 

Always load anonymised or synthetic phantom acquisitions. 

Loads an uncompressed DICOM image (Explicit or Implicit VR Little Endian) and lets you place ROIs interactively on the phantom: a resizable uniform-region ROI for the noise analysis, and six ROIs (five aluminium steps + background) for the PV-to-mmAl calibration. 

Computes the low-contrast detectability curve via the statistical method (Chao et al.): pixel-value thresholds from the standard deviation of sub-ROI means (3.29σ, 95% two one-sided), converted to aluminium thickness through a linear fit on the step measurements. Results are shown as a chart and table, with Excel export.

Note: this tool requires a desktop browser with a mouse. ROI placement, zoom and pan are not optimised for touchscreens. It will not work properly on smartphone or tablet.