Enhanced reduction of color blending by a gamma-based image boosting method for optical see-through displays

Abstract

In augmented reality scenarios, virtual images observed through optical see-through head-mounted displays (OST-HMDs) contain color distortion due to external light. This phenomenon is known as background color blending. Existing methods to overcome this issue based on background subtraction compensation produce negative values when the background is brighter than the virtual image. In such cases, virtual images lose their structural integrity and image quality deteriorates. To solve this problem, we propose a gamma-curve-based virtual image boosting method. We derive a measurement model to determine the appropriate gamma value for virtual image boosting using empirical data obtained from extensive real experiments featuring a variety of virtual images and backgrounds. These diverse experiments prove that the proposed boosting method can effectively improve background color blending. With diverse combinations of natural virtual images and bright backgrounds, including even colorful backgrounds featuring high color saturation, the proposed boosting method produces visually pleasing image quality.