An efficient front‐to‐back depth‐buffer algorithm for real‐time rendering of partially occluded game objects in absence of GPU
Shivendra Shivani · Concurrency and Computation Practice and Experience · 2022
Summary Real‐time rendering of visual scenes is a key requirement in many emerging fields like augmented and virtual reality (AR/VR), computer graphics, and so forth. Latency in rendering the scenes due to massive shading calculations and complete rasterization of view windows is undesirable in real‐time AR/VR scenarios. Real‐time determination of occluded and visible objects is one of the significant problems the rendering process faces. The classical Z‐buffer algorithm used for this purpose requires a lot of memory to store depth values, hence maintaining a buffer. Due to scan‐line tendency, its time complexity is also too high. This article presents a new perspective on a fundamental Z‐buffer algorithm. The proposed approach is not entirely a scan‐line method. It is based on divide and conquer strategy, and it also works on the concept of the front‐to‐back concept, where rays are emitted from a viewpoint toward the objects. Based on the proposed algorithm, partially occluded game objects can also be efficiently determined. Only partial shading calculations will be done using the proposed rendering factor calculation to render those objects. It will also speed up the real‐time processing without using any GPU. This way, the proposed approach reduces the time and space complexity of the rendering process.