paper_3d__utonia

Community

One encoder for all point clouds across domains.

AuthorGonglitian
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Enable unified learning and deployment of 3D representations across diverse point cloud domains with a single encoder, reducing model fragmentation and maintenance.

Core Features & Use Cases

  • Unified cross-domain encoding with a five-stage PTv3 backbone
  • Domain-aware data normalization and grid sampling for cross-domain alignment
  • Serialization + sparse convolution for efficient 3D representation learning
  • Transferable features for segmentation, robotics manipulation, and multimodal reasoning
  • Example use: apply the encoder to indoor scans, outdoor LiDAR, and CAD models to obtain a shared feature space

Quick Start

Install the Utonia encoder and run a sample inference on indoor data using the standard transform.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: paper_3d__utonia
Download link: https://github.com/Gonglitian/agent-skills/archive/main.zip#paper-3d-utonia

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.