Locality-Aware Density Control for Efficient Gaussian-based Image Representation
作者: Jiacong Chen, Qingyu Mao, Xiandong Meng, Shuai Liu, Chao Li, Fanyang Meng, Youneng Bao, Yongsheng Liang
分类: cs.CV
发布日期: 2026-07-20
备注: Accepted by ACMMM 2026
🔗 代码/项目: GITHUB
💡 一句话要点
提出局部感知密度控制以解决高斯图像表示效率问题
🎯 匹配领域: 支柱三:空间感知与语义 (Perception & Semantics)
关键词: 高斯图像表示 密度控制 图像重建 计算机视觉 深度学习 图像压缩 实时渲染
📋 核心要点
- 现有的高斯图像表示方法在优化过程中未能有效分配高斯容量,导致重建内容不足和冗余保留。
- 本文提出局部感知密度控制框架(LocoADC),通过结合局部连续性和相似性来优化高斯分配和合并。
- 在多种数据集上的实验表明,LocoADC在高斯分配的有效性上显著优于多个基线,提升了图像重建质量。
📝 摘要(中文)
2D高斯点云表示因其明确的公式、快速的光栅化和良好的解码效率而受到关注。然而,现有方法在优化过程中未能有效分配高斯容量,导致重建内容的不足和冗余保留。为此,本文提出局部感知密度控制框架(LocoADC),通过区域高斯密度化和相似性驱动的高斯合并策略,结合局部颜色一致性约束,显著提高高斯容量的利用率。实验结果表明,LocoADC在多个基线模型上均有显著提升,在CLIC数据集上相较于GI模型获得了2.93 dB的PSNR增益。
🔬 方法详解
问题定义:本文旨在解决现有高斯图像表示方法在优化过程中高斯容量分配不均的问题,导致重建内容不足和冗余高斯的保留。
核心思路:提出局部感知密度控制框架(LocoADC),利用局部连续性和相似性来优化高斯的分配和合并,从而提高高斯容量的利用率。
技术框架:LocoADC框架包括区域高斯密度化(RGD)和相似性驱动的高斯合并(SDGM)两个主要模块,结合局部颜色一致性约束,确保合并的可靠性。
关键创新:LocoADC的创新在于同时考虑了图像空间中的重建误差连续性和高斯空间中相邻高斯的相似性,从而有效解决了高斯容量的冗余和不足问题。
关键设计:在设计中,RGD模块负责根据重建误差调整高斯的分配,而SDGM模块则通过相似性消除冗余高斯,局部颜色一致性约束确保合并后的高斯在颜色上保持一致。
🖼️ 关键图片
📊 实验亮点
在多项实验中,LocoADC在CLIC数据集上相较于GI模型获得了2.93 dB的PSNR增益,显示出其在高斯分配有效性上的显著提升,且在多个基线模型上均表现出一致的改进效果。
🎯 应用场景
该研究的潜在应用领域包括计算机视觉中的图像重建、图像压缩和实时渲染等。通过提高高斯图像表示的效率,LocoADC能够在图像处理和生成任务中提供更高的质量和更快的速度,具有重要的实际价值和未来影响。
📄 摘要(原文)
2D Gaussian Splatting is an attractive direction for image representation due to its explicit formulation, fast rasterization, and favorable decoding efficiency. The representation quality of this paradigm depends on the proper allocation of Gaussian capacity to the demanding regions. However, existing methods fail to allocate Gaussian capacity efficiently during optimization: under-reconstructed content is often refined in a fragmented pixel-wise manner, while neighboring optimized Gaussians with similar attributes are redundantly retained. This inefficiency motivates the need for a density control framework that jointly addresses insufficient allocation in under-reconstructed regions and redundant allocation in over-reconstructed regions. Our key insight is that this framework should exploit two complementary forms of locality: the local continuity of reconstruction errors in image space for improved Gaussian allocation, and the local similarity of neighboring Gaussians in Gaussian space for redundant elimination. Based on this insight, we propose Locality-Aware Density Control (LocoADC), a plug-and-play framework that improves Gaussian capacity utilization through Region-wise Gaussian Densification (RGD) and Similarity-Driven Gaussian Merging (SDGM) strategies, together with a local color consistency constraint for more reliable merging. Extensive experiments on diverse datasets show that LocoADC consistently improves multiple baselines by enabling more effective local Gaussian allocation, including a 2.93 dB PSNR gain over GI on the CLIC dataset under the same 30k Gaussian budget. Code is available at: \textit{https://github.com/ChenJiaCong-1005/LocoADC}.