AnySplat:基於無約束視角的即時3D高斯潑濺渲染
AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Views
May 29, 2025
作者: Lihan Jiang, Yucheng Mao, Linning Xu, Tao Lu, Kerui Ren, Yichen Jin, Xudong Xu, Mulin Yu, Jiangmiao Pang, Feng Zhao, Dahua Lin, Bo Dai
cs.AI
摘要
我們介紹了AnySplat,這是一個用於從未校準圖像集中進行新視角合成的前饋網絡。與傳統需要已知相機姿態和每場景優化的神經渲染流程,或是近期在密集視角下計算負擔沉重的前饋方法不同,我們的模型一次性預測所有內容。單次前向傳播即可生成一組3D高斯基元,這些基元編碼了場景的幾何與外觀,以及每張輸入圖像對應的相機內外參數。這種統一設計能夠輕鬆擴展至隨意拍攝的多視角數據集,無需任何姿態註釋。在廣泛的零樣本評估中,AnySplat在稀疏和密集視角場景下均能媲美已知姿態基線的質量,並超越了現有的無姿態方法。此外,與基於優化的神經場相比,它大幅降低了渲染延遲,使得在無約束拍攝環境下實現實時新視角合成成為可能。項目頁面:https://city-super.github.io/anysplat/
English
We introduce AnySplat, a feed forward network for novel view synthesis from
uncalibrated image collections. In contrast to traditional neural rendering
pipelines that demand known camera poses and per scene optimization, or recent
feed forward methods that buckle under the computational weight of dense views,
our model predicts everything in one shot. A single forward pass yields a set
of 3D Gaussian primitives encoding both scene geometry and appearance, and the
corresponding camera intrinsics and extrinsics for each input image. This
unified design scales effortlessly to casually captured, multi view datasets
without any pose annotations. In extensive zero shot evaluations, AnySplat
matches the quality of pose aware baselines in both sparse and dense view
scenarios while surpassing existing pose free approaches. Moreover, it greatly
reduce rendering latency compared to optimization based neural fields, bringing
real time novel view synthesis within reach for unconstrained capture
settings.Project page: https://city-super.github.io/anysplat/