// the find
krahets/awesome-mvs
A curated list of tutorials, papers, software related to multi-view stereo.
A reading list of multi-view stereo papers, surveys, benchmarks, and open-source tools, organized by reconstruction approach (depth-map, volumetric, neural implicit). It's for grad students or researchers getting oriented in MVS/3D reconstruction, not for anyone looking for code to run.
The taxonomy is genuinely useful — splitting geometry-based vs. supervised vs. unsupervised learning methods under each representation type gives you a mental map of the field that most paper lists don't bother with. The open-source and commercial software tables include license info, which saves a trip to each repo. It links out to sibling lists (Awesome-Learning-MVS, Awesome-PatchMatch-MVS) instead of duplicating them.
Last commit is August 2022, so three years of the field are missing — nothing on 3D Gaussian Splatting or the newer NeRF/MVS crossover work that dominates current reconstruction research. It's a single README with no scripts, no code, nothing to actually run. Heavy overlap with the other awesome-MVS lists it links to, so it's not clear this one adds anything the others don't already have.