1：Experimental Design and Method Selection:

The study characterized a non-contact imaging scintillator-based dosimetry system for TSET. Scintillator dosimeters were custom-manufactured and attached to phantoms for irradiation. The imaging setup used a time-gated intensified CMOS camera synchronized with the linac. A dose estimation model was developed using MATLAB to fit scintillator emission data to compute surface dose.

2：Sample Selection and Data Sources:

Disc-shaped plastic scintillators (EJ-212) of various thicknesses (0.65–3.13 mm) and diameters (5–30 mm) were used. Phantoms included flat-faced setups and custom containers with water for temperature studies. Reference doses were measured using OSLDs and an ionization chamber.

3：65–13 mm) and diameters (5–30 mm) were used. Phantoms included flat-faced setups and custom containers with water for temperature studies. Reference doses were measured using OSLDs and an ionization chamber. List of Experimental Equipment and Materials:

3. List of Experimental Equipment and Materials: Equipment included a Varian 2100 CD linac, C-Dose intensified CMOS camera, Ocean Optics USB4000 Spectrometer, Fisherbrand 150415C thermometer, PTW 23342 ionization chamber, and nanoDot OSLDs. Materials included EJ-212 scintillators, EJ-510 reflective paint, solid water blocks, and hot plates.

4：Experimental Procedures and Operational Workflow:

Scintillators were irradiated under TSET conditions (6 MeV electrons, 3 m SSD). Imaging was performed with time-gated acquisition, background subtraction, and filtering. Specific tests included temperature variation (10–40°C), emission spectra measurement, dose rate variation (100–1000 MU/min), radiation damage (up to 15,000 Gy), thickness and diameter effects, and scintillator-to-scintillator variation. Data were analyzed using custom MATLAB algorithms.

5：Data Analysis Methods:

Data analysis involved fitting a scintillator model function to image ROIs using trust-region reflective optimization in MATLAB. Linear regression was used for calibration and relationships (e.g., output vs. thickness). Statistical measures included mean, standard deviation, and percent differences.