Noname manuscript No. (will be inserted by the editor) Rasterization-based Progressive Photon Mapping Iordanis Evangelou · Georgios Papaioannou · Konstantinos Vardis · Andreas A. Vasilakis Abstract Ray tracing on the GPU has been syner- gistically operating alongside rasterization in interac- tive rendering engines for some time now, in order to accurately capture certain illumination effects. In the same spirit, in this paper, we propose an implemen- tation of Progressive Photon Mapping entirely on the rasterization pipeline, which is agnostic to the specific GPU architecture, in order to synthesise images at in- teractive rates. While any GPU ray tracing architecture can be used for photon mapping, performing ray traver- sal in image space minimises acceleration data struc- ture construction time and supports arbitrarily complex and fully dynamic geometry. Furthermore, this strategy maximises data structure reuse by encompassing raster- ization, ray tracing and photon gathering tasks in a sin- gle data structure. Both eye and light paths of arbitrary depth are traced on multi-view deep G-buffers and pho- ton flux is gathered by a properly adapted multi-view photon splatting. In contrast to previous methods ex- ploiting rasterization to some extent, due to our novel indirect photon splatting approach, any event combina- tion present in photon mapping is captured. We evalu- ate our method using typical test scenes and scenar- ios for photon mapping methods and show how our approach outperforms typical GPU-based progressive photon mapping. Keywords Photon mapping · Rasterization · Ray tracing I. Evangelou · G. Papaioannou · K. Vardis · A. A. Vasilakis Department of Informatics, Athens University of Economics & Business, Greece Fig. 1 Converged example of our rasterization-based pro- gressive photon mapping method, using depth 3+3 (light + camera). Iteration time (1M pixel samples): 50ms on an NVIDIA RTX 2080 Ti. 1 Introduction Photon Mapping [7,8] is a well-known two-stage ap- proximation to bidirectional path tracing, where light- carrying paths or photons deposit and cache the carried flux on non-specular surfaces, pre-multiplied with the light path throughput. A data structure, the photon map, is responsible for the storage and fast indexing of these particles. Subsequently, for multiple paths traced from the camera, the contribution of photons to hit