ShapeNet
PulseAugur coverage of ShapeNet — every cluster mentioning ShapeNet across labs, papers, and developer communities, ranked by signal.
5 day(s) with sentiment data
-
New Hierarchical Flow Matching method generates 3D point clouds
Researchers have introduced Hierarchical Flow Matching (HFM), a novel method for generating 3D point clouds. HFM addresses limitations in existing flow-based and diffusion models by employing a two-level approach that c…
-
New MiSS Framework Explains 3D Point Cloud Classifier Decisions
Researchers have developed MiSS, a novel framework for explaining the decisions of 3D point cloud classifiers. This black-box system uses perturbation-relative sufficiency reasoning to identify minimal sufficient coalit…
-
3D generative models fail symmetry tests, new paper reveals
A new research paper titled "Symmetry Matters: Auditing and Symmetrizing 3D Generative Models" highlights a significant gap in current 3D generative models: their failure to consistently preserve symmetry in generated o…
-
Sudu Technology unveils embodied AI platform with advanced skill acquisition · 1 source tracked
Sudu Technology, a year-old startup, has showcased its embodied AI platform, Sudo R1, demonstrating advanced capabilities in object manipulation and task execution. Founded by Professor Su Hao, a prominent figure in 3D …
-
New AI method generates 3D furniture codes without labels
Researchers have developed a new method for generating 3D furniture models without requiring explicit labels or pose annotations. By using a Finite Scalar Quantization autoencoder trained on the 3D-FUTURE dataset, the s…
-
3DMPE method reconstructs 3D point clouds from partial multi-view projections
Researchers have introduced 3DMPE, a novel training-free method for reconstructing 3D point clouds from multiple 2D projections. This optimization-based approach handles scenarios where different views capture varying s…
-
GenSP framework learns consistent spherical parameterizations for 3D shapes
Researchers have developed GenSP, a novel framework for creating consistent spherical parameterizations across various 3D shapes. Unlike previous methods that optimize each shape independently, GenSP learns a neural gen…
-
New framework enhances 3D generative model interpretability
Researchers have developed a framework called 3D-CBM to enhance interpretability in 3D generative models by integrating Concept Bottleneck Models. This approach aims to bridge the semantic gap in deep geometric learning…
-
New method learns clean 3D neural fields from noisy data
Researchers have developed a new method called NoiseSDF2NoiseSDF to improve the reconstruction of 3D neural fields from noisy point cloud data. This technique extends the Noise2Noise paradigm from 2D images to 3D, enabl…
-
New AI Models Advance 3D Shape Completion and Depth Estimation
Researchers have introduced several new models for 3D shape completion and depth estimation. The Large Depth Completion Model (LDCM) uses a transformer to generate dense depth maps from sparse observations, outperformin…
-
EvObj advances unsupervised 3D instance segmentation with domain adaptation
Researchers have developed EvObj, a novel approach for unsupervised 3D instance segmentation that overcomes the domain gap between synthetic and real-world data. The method employs an object discerning module to adapt o…
-
New Orbit-Space Particle Flow Matching framework enhances generative modeling
Researchers have introduced Orbit-Space Geometric Probability Paths (OGPP), a novel framework for generative modeling of particle systems. This approach addresses challenges related to particle permutation symmetries an…
-
RETO Transformer operator enhances automotive aerodynamics prediction with RoPE
Researchers have introduced RETO, a novel rotary-enhanced transformer operator designed to improve the prediction of automotive aerodynamics. This new model incorporates a dual-stage spatial awareness mechanism, utilizi…