Methods for estimating full width at half maximum

Abstract Background: Measurement of full width at half maximum (FWHM) is used in multiple important applications related to the evaluation of the image resolution and the performance assessment of different scanner devices especially in positron emission tomography (PET) imaging, but little research has been performed about different methods for estimating FWHM. Materials and methods : We study seven different methods for estimating FWHM, including measuring FWHM according to its definition, four methods based on estimation of the standard deviation, the standard method of National Electrical Manufacturers Association (NEMA), and a new method utilizing an optimization algorithm to fit a Gaussian curve to the data. We compare their performance with several simulations. We also apply these techniques on two real data sets created by PET images of an $$^{18}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mmultiscripts> <mml:mrow/> <mml:mrow/> <mml:mn>18</mml:mn> </mml:mmultiscripts> </mml:math> F-fluorodeoxyglucose resolution phantom imaged with High Resolution Research Tomograph concept camera and Biograph Vision Quadra and a sodium-22 point source imaged with RAYCAN X5 PET/CT system. Results and conclusion : According to our results from both simulated and real data, the most accurate methods are the first method based on the definition of FWHM and the NEMA standard method. In particular, both of these methods work in a reliable way in real data experiments, even when there is a very limited number of points available for FWHM computation.