EcoSplat: Efficiency-controllable Feed-forward 3D Gaussian Splatting from Multi-view Images

TL;DR

EcoSplat introduces a novel, efficiency-controllable 3D Gaussian Splatting framework that adaptively predicts scene representations with explicit primitive count control, improving efficiency and flexibility in novel view synthesis.

Contribution

It is the first feed-forward 3D Gaussian Splatting method with explicit primitive count control, using a two-stage training process for adaptive scene representation prediction.

Findings

Outperforms state-of-the-art methods under primitive count constraints

Robust across multiple dense-view settings

Enables flexible downstream rendering tasks

Abstract

Feed-forward 3D Gaussian Splatting (3DGS) enables efficient one-pass scene reconstruction, providing 3D representations for novel view synthesis without per-scene optimization. However, existing methods typically predict pixel-aligned primitives per-view, producing an excessive number of primitives in dense-view settings and offering no explicit control over the number of predicted Gaussians. To address this, we propose EcoSplat, the first efficiency-controllable feed-forward 3DGS framework that adaptively predicts the 3D representation for any given target primitive count at inference time. EcoSplat adopts a two-stage optimization process. The first stage is Pixel-aligned Gaussian Training (PGT) where our model learns initial primitive prediction. The second stage is Importance-aware Gaussian Finetuning (IGF) stage where our model learns rank primitives and adaptively adjust their parameters based on the target primitive count. Extensive experiments across multiple dense-view settings show that EcoSplat is robust and outperforms state-of-the-art methods under strict primitive-count constraints, making it well-suited for flexible downstream rendering tasks.