AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Views
This enables real-time novel view synthesis for casually captured, multi-view datasets, addressing a bottleneck in unconstrained settings, though it is incremental as it builds on existing 3D Gaussian splatting methods.
The paper tackles novel view synthesis from unconstrained image collections without camera poses, introducing AnySplat, a feed-forward network that predicts 3D Gaussian primitives and camera parameters in a single pass, matching pose-aware baselines in quality and reducing rendering latency for real-time performance.
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/