Bae Joungeun, Yoo Hoon
Department of Computer Science, Sangmyung University, 20 Hongjimoon-2gil, Seoul 030031, Korea.
Department of Electronics Engineering, Sangmyung University, 20 Hongjimoon-2gil, Seoul 030031, Korea.
Sensors (Basel). 2020 Aug 25;20(17):4795. doi: 10.3390/s20174795.
This paper describes the image enhancement of a computational integral imaging reconstruction method via reconstructing a four-dimensional (4-D) image structure. A computational reconstruction method for high-resolution three-dimensional (3-D) images is highly required in 3-D applications such as 3-D visualization and 3-D object recognition. To improve the visual quality of reconstructed images, we introduce an adjustable parameter to produce a group of 3-D images from a single elemental image array. The adjustable parameter controls overlapping in back projection with a transformation of cropping and translating elemental images. It turns out that the new parameter is an independent parameter from the reconstruction position to reconstruct a 4-D image structure with four axes of , , , and . The 4-D image structure of the proposed method provides more visual information than existing methods. Computer simulations and optical experiments are carried out to show the feasibility of the proposed method. The results indicate that our method enhances the image quality of 3-D images by providing a 4-D image structure with the adjustable parameter.
本文通过重建四维（4-D）图像结构描述了一种计算积分成像重建方法的图像增强。在三维可视化和三维目标识别等三维应用中，对高分辨率三维（3-D）图像的计算重建方法有很高的要求。为了提高重建图像的视觉质量，我们引入一个可调参数，从单个基元图像阵列生成一组三维图像。该可调参数通过裁剪和平移基元图像的变换来控制反投影中的重叠。结果表明，新参数是一个与重建位置无关的独立参数，用于重建具有x、y、z和t四个轴的四维图像结构。所提方法的四维图像结构比现有方法提供了更多的视觉信息。进行了计算机模拟和光学实验以证明所提方法的可行性。结果表明，我们的方法通过提供具有可调参数的四维图像结构来提高三维图像的质量。
