偏移量(计算机科学)
计算机科学
编码(内存)
算法
计算机图形学(图像)
人工智能
程序设计语言
作者
Peiren Wang,Zhengyang Li,Xiaozheng Wang,Wei Wang,Qin Liu
摘要
In the realm of rendering algorithms predicated upon depth information, the escalating image resolution poses formidable challenges, culminating in protracted rendering cycles requisite for the synthesis of stereoscopic display images. This impediment renders the attainment of real‐time rendering a formidable undertaking. Therefore, an efficient 3D image encoding algorithm based on depth offset mapping is proposed. Using a two‐dimensional color image as a reference, the algorithm utilizes its corresponding depth map to obtain depth information for calculating the offset of each sub‐pixel. By directly applying the geometric relationship between the display and viewing positions and the principle of reversible light paths, three‐dimensional image synthesis rendering is performed. This allows each eye of the viewer to see a parallax image composed of sub‐pixels with different offset values, creating a sense of depth perception. In addition to ensuring display efficacy, the proposed methodology eschews conventional paradigms involving the generation of copious virtual viewpoints, thereby mitigating the memory demands imposed upon hardware systems and fostering the expeditious integration and rendering of three‐dimensional images within a systemic framework.
科研通智能强力驱动
Strongly Powered by AbleSci AI