SGI: Structured 2D Gaussians for Efficient and Compact Large Image Representation

2026-03-08

AI Summary

This work addresses the limitations of conventional 2D Gaussian splatting in high-resolution image representation, which suffers from parameter redundancy, slow convergence, and large storage overhead. The authors propose the Structured Gaussian Image (SGI) framework, which constructs multi-scale local neighborhoods anchored by seed points and employs a lightweight MLP to generate spatially coherent implicit 2D neural Gaussians. A coarse-to-fine multi-scale optimization strategy is introduced to enhance efficiency. The method further enables entropy-based seed-level compression, achieving up to 7.5× higher compression ratios and 1.6× faster optimization compared to existing non-quantized approaches, while maintaining or even improving image fidelity. Relative to quantized baselines, it attains 1.6× better compression and 6.5× acceleration.

Technology Category

• Computer Vision: Representation Learning for Vision
• Machine Learning: Representation Learning
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

2D Gaussian Splatting has emerged as a novel image representation technique that can support efficient rendering on low-end devices. However, scaling to high-resolution images requires optimizing and storing millions of unstructured Gaussian primitives independently, leading to slow convergence and redundant parameters. To address this, we propose Structured Gaussian Image (SGI), a compact and efficient framework for representing high-resolution images. SGI decomposes a complex image into multi-scale local spaces defined by a set of seeds. Each seed corresponds to a spatially coherent region and, together with lightweight multi-layer perceptrons (MLPs), generates structured implicit 2D neural Gaussians. This seed-based formulation imposes structural regularity on otherwise unstructured Gaussian primitives, which facilitates entropy-based compression at the seed level to reduce the total storage. However, optimizing seed parameters directly on high-resolution images is a challenging and non-trivial task. Therefore, we designed a multi-scale fitting strategy that refines the seed representation in a coarse-to-fine manner, substantially accelerating convergence. Quantitative and qualitative evaluations demonstrate that SGI achieves up to 7.5x compression over prior non-quantized 2D Gaussian methods and 1.6x over quantized ones, while also delivering 1.6x and 6.5x faster optimization, respectively, without degrading, and often improving, image fidelity. Code is available at https://github.com/zx-pan/SGI.

Problem

• 2D Gaussian Splatting
• high-resolution image representation
• unstructured Gaussian primitives
• parameter redundancy
• slow convergence

Innovation

• Structured Gaussian
• 2D Gaussian Splatting
• Multi-scale Representation
• Entropy-based Compression
• Coarse-to-fine Optimization