Real-Time Rendering - Science topic

I recently found a dissertation about lightmap compression: https://www.diva-portal.org/smash/get/diva2:844146/FULLTEXT01.pdf

Are there other papers about this topic? Thanks.

Google's Bilateral Guided Upsampling (2016) proposed an idea that upsampling could be guided by a reference image to maintain high-frequency information like edges. This reminds me of G-buffers in real-time rendering.

Does state-of-the-art super-resolution algorithms in real-time renderings, such as Nvidia's DLSS and AMD's FSR, use a similar idea? How do they exploit G-buffers to aid the upsampling?

I'm trying to segment the mitochondrial fusion process and hence, would like to track the temporal difference between the inner and outer membrane fusion. I have a Tim23-RFP for the inner membrane. Any good suggestions for marking the intermembrane space would be welcome.

Best,

Roy

I am trying to detect 3d model from live video stream.Model should be detected by any face.how can  do that?

I'm analysing ER-mitochondria contact sites but I can not find a detailed protocol to obtain 3D reconstructed and surface rendered images with ImageJ.Thank you

Graphics renders 3D scenes from render lists and scene descriptions (a render language) and Computer Vision seems to be advancing to where one can imagine a camera and co-processor (like a GPU) that would generate an object list and scene description from continuously captured camera frames (depth map, segmented/recognized objects, textures, lighting). So in some sense making CV is just the inverse of graphics rendering. The 3D scanner and printer are already a reality. At what point does this fuse so that capture to render is indistinguishable from a digital video stream?