HCC: Small: Individualized Inverse-Blurring and Aberration Compensated Displays for Personalized Vision Correction with Applications for Mobile Devices

本研究关注的是眼睛的屈光因素和在将光聚焦到视网膜上的过程中产生的错误或伪影。随着移动手持设备的不断普及，一个限制因素可能会越来越多地阻碍它们在相当大的人群中被采用，那就是视力问题的普遍存在。这些设备的有效使用取决于必须与小区域交互的用户的合理视觉表现。这对老年用户来说尤其成问题，因为随着年龄的增长，他们面临着越来越多的视力疾病。但即使是年轻人，也有证据表明近视的发病率正在上升，尤其是在亚洲人群中。此外，一些视觉障碍涉及高阶光学像差（有时被称为“不规则散光”），这是不可能用眼镜镜片纠正的。在之前的工作中，PI开发了视觉逼真渲染（VRR），通过测量他或她的光学系统来模拟个人的视觉系统。考虑到这些相同的光学测量值，PI在当前项目中的目标是通过算法和数字而不是光学来实现视力矫正；也就是说，给定一个通过眼镜矫正屈光不正或具有高阶光学像差的用户，计算一个个性化的“反向模糊”变换，应用于一个清晰的图像，这样当这个人看到变换后的图像时，反向模糊被他或她的视力的光学像差抵消，这个模糊的图像版本在这个人看来是清晰的。逆模糊处理的问题是，它往往会产生比原始图像动态范围大得多的图像（由于模糊核中的弱频率响应，并且除以弱响应会产生较大的值）。在预反卷积图像中通常有许多负像素和一个亮点。预反卷积图像的计算涉及使用反滤波或空间域求解器。然而，这种情况与将图像去模糊作为后处理的情况根本不同；由于模糊卷积是最后一步，因此存在无法通过添加先验知识来恢复的频率信息损失。PI在这个项目中的方法是通过使用高动态范围显示系统来解决预反卷积图像的大动态范围问题，并通过多层显示的概念来解决频率信息的丢失问题，即使在模糊之后也不会丢失任何频率内容。更广泛的影响：眼镜不能纠正许多患有某些类型角膜病变或经历角膜屈光手术（如LASIK和PRK）副作用的患者的视觉系统中出现的高阶光学像差。如果成功，本研究将显著影响视力矫正技术，为各种新的显示算法和设备奠定基础，这些算法和设备将超越这一限制，为低阶或高阶光学像差相关的视力问题患者提供视力矫正。私家侦探努力让本科生和少数民族参与他的研究。他为各个层次的学生提供独立学习，包括本科生参加研究小组会议，并为大一和大二学生提供相关主题的研讨会。他与黑人研究生工程和科学学生以及拉丁裔工程和科学研究生协会密切合作，并监督伯克利夏季工程研究本科项目的学生，该项目吸引了来自全国各地的少数族裔在夏季到伯克利做研究。

This research is concerned with the refractive elements of the eye and the errors or artifacts produced in the process of focusing light onto the retina. As the popularity of mobile hand-held devices continues to grow, a limiting factor that may emerge as an increasing impediment to their adoption among sizeable segments of the population is the prevalence of vision problems. Effective use of these devices is predicated upon reasonable visual performance by a user who must interact with a small area. This is particularly problematic for the population of older users, who face increasing incidence of vision ailments as they age. But even for younger people, there is evidence that the prevalence of myopia is increasing, especially in Asian populations. Furthermore, some visual impairments involve higher order optical aberrations (sometimes referred to as "irregular astigmatism"), which are impossible to correct with spectacle lenses.In prior work, the PI developed Vision-Realistic Rendering (VRR) to simulate an individual's vision system from measuring his or her optical system. Given these same optical measurements for that individual, the PI's goal in the current project is to achieve vision correction algorithmically and digitally rather than optically; that is, given a user with refractive error corrected by spectacles or with high order optical aberrations, compute an individualized "inverse blur" transformation to be applied to a sharp image such that when the resulting transformed image is then viewed by this individual, the inverse blur is canceled by the optical aberrations of his or her vision and this blurred version of the image appears in sharp focus to this individual.The problem with the inverse blurring process is that it tends to produce an image whose dynamic range is much larger than that of the original image (due to a weak frequency response in the blurring kernel, and division by weak response creates large values). There are usually many negative pixels and a bright spot in the pre-deconvolved image. Computation of the pre-deconvolved image involves using inverse-filtering or a spatial domain solver. However, the situation is fundamentally different from that of performing the image de-blurring as a post-process; since the blurring convolution is the final step, there is a loss of frequency information that cannot be recovered by adding prior knowledge. The PI's approach in this project is to address the large dynamic range of the pre-deconvolved image by using a high dynamic range display system, and the loss of frequency information by the concept of a multi-layered display that does not lose any frequency content even after the blurring.Broader Impacts: Eyeglasses cannot correct higher order optical aberrations that arise in the vision system of many patients who have certain types of corneal pathologies or who experience side effects of corneal refractive surgeries (such as LASIK and PRK). If successful, this research will significantly impact vision correction technology by laying the foundations for a variety of new display algorithms and devices which transcend this limitation and provide vision correction for patients whose vision problems are related to either low or higher order optical aberrations. The PI makes a concerted effort to involve undergraduate students and minorities in his research. He offers independent study for students at all levels, includes undergraduate students in research group meetings, and offers freshman and sophomore seminars on related topics. He works closely with the Black Graduate Engineering and Science Students and Latino Association of Graduate Students in Engineering and Science, and supervises students in the Summer Undergraduate Program in Engineering Research at Berkeley which brings underrepresented minorities from around the country to do research at Berkeley during the summer.