17.3: Color Correction in TFTLCD Displays for Compensation of Color Dependency with the Viewing Angle

Abstract

Most today's twisted nematic TFTLCD displays show a color dependency with the viewing angle. This paper describes a method for compensation of the pixel color variations in these displays relative to a single viewer position. The color correction can be recomputed automatically as the viewer position changes, as long as the position is known. The method determines the color correction required for each pixel of the screen, such that a single viewer for a given position can see the color unaffected by the viewing angle differences to the screen. Color is measured in few points over the screen relative to a viewer position, and then an interpolation procedure is used to compute the color correction for each intermediate pixel, between the measured positions. The measurements for different positions of the viewer are used in a second interpolation procedure for computing color correction for intermediate position of the viewer between the measured positions. The method was experimented for portable and desktop displays. It was found that, for relative small viewing angle variations, with acceptable approximation, the color variation with the viewing angle can be modeled as variations of the transfer function. It was found that the proposed method can also be effectively used to compensate for panel uniformity.

Real-time Compensation for Angular-dependent Color Shift in TN LCDs

“…Marcu et al proposed a method of compensating for the color shift at different locations [1]. Li et al proposed an idea of dynamically compensating for the color shift according to the viewer's position [2].…”

P‐27: Adaptive Display Color Correction Based on Real‐time Viewing Angle Estimation

“…Existing approaches either attempt to improve the color characteristics for general wide viewing angles at the expense of the overall image quality (thus cannot achieve optimal results for the real viewing angle of the user)(e.g., [1]), or attempt to do color correction with limited viewing-angle-adaptation through predetermined assumptions on viewing angles (which could be completely invalid in most applications since the user may freely change his/her position with respect to the display) (e.g. [2]).…”

“…For Case I, one could assuming that the viewer's normal viewing angle is along the axis at the center of the display so that one can use, for example, nine different gammas for nine sub-regions, as illustrated in Figure 5. This is the method proposed in [2], but it works only when the viewer stays at the fixed position with respect to the display since all the nine gamma correction curves have to be pre-determined based on the viewer's position. With the proposed the method, after the user's viewing angle with respect to the center of the display is estimated, we can further compute the respective viewing angles for other sub-regions (or even smaller sub-regions) by estimating the user's distance to the display (the estimation can be done, for example, by face detection with the assumption of the normal size of a typical face, or with a stereo camera).…”

Advances in artificial intelligence for the colour industry

“…The mapping of colour monitor RBG values onto CMYK printing coordinates has been discussed and the following advantages of the neural system over the traditional approach proposed [31]: (a) The colour conversion is fast, since the required computations are trivial (b) The network can continue to learn after the initial training process and thus offers potential adaptation to more complex phenomena.…”