NoPoSplat

Visit Tool

NoPoSplat is a Content & Design tool that predicts 3D Gaussians from sparse, unposed images. It enables high-quality novel view synthesis and accurate pose estimation for 3D reconstruction.

Claim this tool

1View

At a glance

Pricing

Open Source

Free tier

Yes

API

Skill level

Technical

About

What is NoPoSplat?

NoPoSplat is an open-source research project that leverages 3D Gaussian Splats to generate high-quality novel views and perform accurate pose estimation from sparse, unposed images. Developed based on PyTorch, it allows users to predict 3D Gaussians in a canonical space, addressing challenges in 3D reconstruction where camera poses are unknown. The tool provides pre-trained checkpoints for various datasets like RealEstate10K and ACID, and includes detailed instructions for installation, training, and evaluation of both novel view synthesis and pose estimation. It's ideal for researchers and developers working on advanced computer vision tasks.

Best used for

Ideal for developers and data scientists who need to perform 3D reconstruction from sparse, unposed images, synthesize novel views, and accurately estimate camera poses. Especially valuable for research and development in computer vision and 3D graphics, offering a robust solution for challenging unposed image datasets.

Common actions

predict 3D Gaussians

synthesize novel views

estimate camera pose

reconstruct 3D scenes

workflowsdeepfakelow-code/no-codecollaborationautomated workflowopen-sourcegithub copilotface swapping"AI Agents"

Capabilities

Key features

3D Gaussian prediction
Novel view synthesis
Accurate pose estimation
Sparse unposed images
Pre-trained checkpoints

Target Audience

developerdata scientist

Integrations

Not yet documented

Pricing & Plans

Open Source

Free

FAQs

What are the hardware requirements for running NoPoSplat?

NoPoSplat requires Python 3.10+, PyTorch 2.1.2, and CUDA 11.8. For training, the default configuration needs 8 GPUs with 16 batch size each (>=80GB memory). A configuration for 1 A6000 GPU (48GB memory) is also available.

Which datasets does NoPoSplat support for training and evaluation?

NoPoSplat supports various datasets including RealEstate10K, ACID, DL3DV, DTU, and ScanNet++. Pre-trained models are available for combinations of these datasets, and instructions are provided for preparing your own datasets.

Can NoPoSplat be used for both novel view synthesis and pose estimation?

Yes, NoPoSplat is designed to perform both high-quality novel view synthesis and accurate pose estimation from sparse, unposed images. Separate evaluation scripts are provided for testing each of these capabilities.

Trending

Subcategories trending in Content & Design

Image Generation AI Writing Assistants Audio & Music Video Generation Photo Editing Graphic Design

Trending

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce