New method audits 3D shape retrieval, GMSD-HKS leads benchmarks

A new research paper introduces Diffused Geodesic Moments (DGM), a novel method for evaluating training-free 3D shape descriptors. The study reframes descriptor evaluation as a protocol audit, highlighting how local signal design, normalization, and aggregation choices significantly impact retrieval scores. Experiments on the FAUST and TOSCA datasets show that an independent Geometric Moment Shape Descriptor baseline using Heat Kernel Signature features (GMSD-HKS) achieves the highest scores, while DGM proves useful for specific applications like sparse solves or non-spectral deployment. AI

RANK_REASON This is a research paper detailing a new methodology and baseline for evaluating 3D shape retrieval. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• Diffused Geodesic Moments
• FAUST
• Geometric Moment Shape Descriptor
• Heat Kernel Signature
• TOSCA
• Wave Kernel Signature

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →