Curve-based Depth Map Image Adjustment for Eyeball Phantom Measurement

Abstract

Light field images have much more information than regular images. One of the advantages is that this processing can produce depth map images. However, there is a problem, the resulting depth map image is inconsistent due to the influence of lighting variations around the object. Therefore it is necessary to make adjustments so that the depth map value matches the actual size. The method proposed in this paper is to adjust the depth map image with the slit lamp guide curve. The object being measured is a ground truth reference object with some defined surface height used for measurement and calibration, and the other object is an eyeball phantom. First, a consumer light field camera is attached to a slit lamp for image acquisition. In the next step, the depth map obtained from the light field image is adjusted to increase the range of pixel values. The third step is the depth calculation of the projected slit light. The last step is the adjustment of the object surface curve based on the depth obtained from the projected slit light. Based on testing on both objects, the proposed method can improve the curvature of the depth map.