REVIEW 15 cited by
Long-LRM: Long-sequence Large Reconstruction Model for Wide-coverage Gaussian Splats
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
Long-LRM: Long-sequence Large Reconstruction Model for Wide-coverage Gaussian Splats
read the original abstract
We propose Long-LRM, a feed-forward 3D Gaussian reconstruction model for instant, high-resolution, 360{\deg} wide-coverage, scene-level reconstruction. Specifically, it takes in 32 input images at a resolution of 960x540 and produces the Gaussian reconstruction in just 1 second on a single A100 GPU. To handle the long sequence of 250K tokens brought by the large input size, Long-LRM features a mixture of the recent Mamba2 blocks and the classical transformer blocks, enhanced by a light-weight token merging module and Gaussian pruning steps that balance between quality and efficiency. We evaluate Long-LRM on the large-scale DL3DV benchmark and Tanks&Temples, demonstrating reconstruction quality comparable to the optimization-based methods while achieving an 800x speedup w.r.t. the optimization-based approaches and an input size at least 60x larger than the previous feed-forward approaches. We conduct extensive ablation studies on our model design choices for both rendering quality and computation efficiency. We also explore Long-LRM's compatibility with other Gaussian variants such as 2D GS, which enhances Long-LRM's ability in geometry reconstruction. Project page: https://arthurhero.github.io/projects/llrm
Forward citations
Cited by 15 Pith papers
-
Stream3D: Sequential Multi-View 3D Generation via Evidential Memory
Stream3D is a training-free method that maintains temporal consistency in 3D generation from monocular streams by dynamically caching a fixed number of informative historical frames using an evidence score.
-
AdaptSplat: Adapting Vision Foundation Models for Feed-Forward 3D Gaussian Splatting
AdaptSplat adds a Frequency-Preserving Adapter to vision foundation models to boost high-frequency fidelity and cross-domain performance in feed-forward 3D Gaussian Splatting.
-
StreamSplat: Streaming Feed-Forward 3D Gaussian Splatting
StreamSplat incrementally builds a 3D Gaussian scene from a causal view stream using a memory-bounded voxel cache, and scales to hundreds of views where fixed-context baselines run out of memory.
-
Lighting-Consistent Object Transfer Across Radiance Fields
Diffusion-based per-view harmonization for lighting-consistent object transfer between 3DGS scenes, using heterogeneous training data and final 3D consolidation.
-
Stream3D: Sequential Multi-View 3D Generation via Evidential Memory
Stream3D is a training-free method that maintains a fixed-size evidential memory of past frames to convert frozen view-conditioned 3D generators into consistent streaming generators.
-
Resolving Representation Ambiguity in Feedforward Novel View Synthesis Transformer via Semantic-Spatial Decoupling
Decouples semantic and spatial tokens in NVS transformers to resolve representation ambiguity, yielding consistent gains with near-zero added latency.
-
Feed-Forward 3D Scene Modeling: A Problem-Driven Perspective
The paper proposes a problem-driven taxonomy for feed-forward 3D scene modeling that groups methods by five core challenges: feature enhancement, geometry awareness, model efficiency, augmentation strategies, and temp...
-
FreeScale: Scaling 3D Scenes via Certainty-Aware Free-View Generation
FreeScale generates scalable high-quality training data for generalizable novel view synthesis by certainty-aware sampling from imperfect scene reconstructions, delivering 2.7 dB PSNR gains on out-of-distribution tests.
-
GeoNVS: Geometry Grounded Video Diffusion for Novel View Synthesis
Feature-space Gaussian Splat Feature Adapter (GS-Adapter) grounds camera-controlled video diffusion in 3D Gaussians, improving geometric consistency and controllability over SEVA and CameraCtrl without retraining geom...
-
Flux4D: Flow-based Unsupervised 4D Reconstruction
Flux4D reconstructs large-scale dynamic 4D scenes unsupervised by predicting moving 3D Gaussians from photometric losses and static regularization when trained across multiple scenes.
-
VolSplat: Rethinking Feed-Forward 3D Gaussian Splatting with Voxel-Aligned Prediction
VolSplat predicts 3D Gaussians from a shared voxel grid instead of from image pixels, reporting large gains in sparse-view novel view synthesis on RealEstate10K, ScanNet, and ACID.
-
Test-Time Training Done Right
Large-chunk online updates during inference let test-time training scale state capacity to 40% of model size and handle contexts up to 1M tokens without custom kernels.
-
AdaptSplat: Adapting Vision Foundation Models for Feed-Forward 3D Gaussian Splatting
AdaptSplat adds a lightweight Frequency-Preserving Adapter to vision foundation models that extracts direction-aware high-frequency priors and integrates them via positional encodings and residual modulation to improv...
-
Long-LRM++: Preserving Fine Details in Feed-Forward Wide-Coverage Reconstruction
Long-LRM++ achieves real-time 14 FPS high-fidelity 360-degree scene reconstruction from 32-64 views by using semi-explicit Gaussians plus a light decoder, matching LaCT quality on DL3DV and improving depth prediction.
-
Hestia: Voxel-Face-Aware Hierarchical Next-Best-View Acquisition for Efficient 3D Reconstruction
Hestia improves generalizable next-best-view planning for 3D reconstruction via hierarchical action search, diverse data, close-greedy strategy, and face-aware voxel design, yielding higher coverage and lower Chamfer ...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.