REVIEW 7 cited by
Instant Neural Graphics Primitives with a Multiresolution Hash Encoding
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
abstract
Neural graphics primitives, parameterized by fully connected neural networks, can be costly to train and evaluate. We reduce this cost with a versatile new input encoding that permits the use of a smaller network without sacrificing quality, thus significantly reducing the number of floating point and memory access operations: a small neural network is augmented by a multiresolution hash table of trainable feature vectors whose values are optimized through stochastic gradient descent. The multiresolution structure allows the network to disambiguate hash collisions, making for a simple architecture that is trivial to parallelize on modern GPUs. We leverage this parallelism by implementing the whole system using fully-fused CUDA kernels with a focus on minimizing wasted bandwidth and compute operations. We achieve a combined speedup of several orders of magnitude, enabling training of high-quality neural graphics primitives in a matter of seconds, and rendering in tens of milliseconds at a resolution of ${1920\!\times\!1080}$.
Forward citations
Cited by 7 Pith papers
-
StreamME: Simplify 3D Gaussian Avatar within Live Stream
StreamME reconstructs an animatable head avatar from a live monocular video in about five minutes by attaching 3D Gaussian points to a tracked face mesh and pruning unimportant points during training.
-
VTGaussian-SLAM: RGBD SLAM for Large Scale Scenes with Splatting View-Tied 3D Gaussians
A new RGBD SLAM representation ties Gaussian positions to depth pixels, leaving only color, radius, and opacity learnable, enabling local-only optimization and higher rendering quality on several benchmarks.
-
Quo Vadis, World Modeling?
An agent-centric reframing of world modeling, replacing physical state prediction with 'information transitions' organized into six proxy functions and three empowerment levels.
-
Construction of Digital Terrain Maps from Multi-view Satellite Imagery using Neural Volume Rendering
Neural terrain maps reconstruct digital elevation models from multi-view satellite imagery alone, reaching near image-resolution accuracy.
-
UAV4D: Dynamic Neural Rendering of Human-Centric UAV Imagery using Gaussian Splatting
UAV4D reconstructs 4D scenes from monocular drone video by fitting a single global scale to align human meshes with the background mesh, then renders with separate Gaussian splats.
-
BayesSDF: Surface-Based Laplacian Uncertainty Estimation for 3D Geometry with Neural Signed Distance Fields
BayesSDF estimates surface uncertainty as aggregated squared color gradients with respect to a hash-encoded deformation field, and reports correlation with depth error on four benchmark scenes.
-
NeurNCD: Novel Class Discovery via Implicit Neural Representation
NeurNCD proposes a NeRF-based framework for novel class discovery in RGB-D scenes, claiming superior mIoU on NYUv2 and Replica, though the presented implementation is internally inconsistent.
Discussion (0). Sign in to comment.