Analysis method of noise power spectrum for medical monochrome liquid crystal displays

Abstract

We investigated methods of analyzing the noise power spectrum (NPS) measurement for medical liquid crystal displays (LCDs). Uniform images displayed on the LCDs were imaged with a high-performance digital camera equipped with a close-up lens, and then the NPSs were calculated from the image data by means of several analysis methods. In a method using the 2D fast Fourier transform (FFT) with a 256 × 256 pixels data segment (basic method), we examined the efficacy of a background trend correction (BTC) and a Hanning windowing process used for reducing the spectral estimation errors in the Fourier analysis. To improve the frequency resolution of the basic method, we examined two 2D FFT methods by using 512 × 512 and 1024 × 1024 pixel segments. In addition, we studied a 1D FFT method with 1024-point 1D noise profiles (1D method). In these three methods, the BTC by a second-order polynomial fit and Hanning windowing were commonly applied. A 3-mega-pixel (MP) and a 5-MP monochrome LCD were employed for evaluating the respective methods. Also, a prototype 5-MP LCD equipped with a new anti-reflection surface-coated panel was compared with the conventional 5-MP LCD. The Hanning windowing process was indispensable for avoiding the spectral leakage errors caused by the pixel structures of the LCD. Sufficient frequency resolution was obtained by the 2D FFT method with the 1024 × 1024 pixels segments and the 1D method. The method which provided the most reliable NPSs was the 1D method, with which the BTC was achieved successfully.