Premium
Resolution and noise measurements of five CRT and LCD medical displays
Author(s) -
Saunders Robert S.,
Samei Ehsan
Publication year - 2006
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.2150777
Subject(s) - liquid crystal display , optics , pixel , nyquist frequency , optical transfer function , luminance , subpixel rendering , cathode ray tube , charge coupled device , image quality , noise (video) , display device , image resolution , image sensor , materials science , physics , artificial intelligence , computer vision , computer science , filter (signal processing) , image (mathematics) , operating system
The performance of soft‐copy displays plays a significant role in the overall image quality of a digital radiographic system. In this work, we discuss methods to characterize the resolution and noise of both cathode ray tube (CRT) and liquid crystal display (LCD) devices. We measured the image quality of five different commercial display devices, representing both CRT and LCD technologies, using a high‐quality charge‐coupled device (CCD) camera. The modulation transfer function (MTF) was calculated using the line technique, correcting for the MTF of the CCD camera and the display pixel size. The normalized noise power spectrum (NPS) was computed from two‐dimensional Fourier analysis of uniform images. To separate the effects of pixel structure from interpixel luminance variations, we created structure‐free images by eliminating the pixel structures of the display device. The NPS was then computed from these structure‐free images to isolate interpixel luminance variations. We found that the MTF of LCDs remained close to the theoretical limit dictated by their inherent pixel size ( 0.85 ± 0.08 at Nyquist frequency), in contrast to the MTF for the two CRT displays, which dropped to 0.15 ± 0.08 at the Nyquist frequency. However, the NPS of LCDs showed significant peaks due to the subpixel structure, while the NPS of CRT displays exhibited a nearly flat power spectrum. After removing the pixel structure, the structured noise peaks for LCDs were eliminated and the overall noise magnitude was significantly reduced. The average total noise‐to‐signal ratio for CRT displays was 6.55 % ± 0.59 % , of which 6.03 % ± 0.24 % was due to interpixel luminance variations, while LCD displays had total noise to signal ratios of 46.1 % ± 5.1 % of which 1.50 % ± 0.41 % were due to interpixel luminance variations. Depending on the extent of the blurring and prewhitening processes of the human visual system, the magnitude of the display noise (including pixel structure) potentially perceived by the observer was reduced to 0.43 % ± 0.01 % (accounting for blurring only) and 0.40 ± 0.01 % (accounting for blurring and prewhitening) for CRTs, and 1.02 % ± 0.22 % (accounting for blurring only) and 0.36 % ± 0.08 % (accounting for blurring and prewhitening) for LCDs.