SASNet: Spatially-Adaptive Sinusoidal Networks for INRs
SIRENs are powerful implicit neural representations, but they are highly sensitive to frequency initialization: low frequencies over-smooth fine detail, while high frequencies inject noise into smooth regions. SASNet addresses this with a fixed frequency embedding paired with jointly learned spatially-adaptive masks that localize each sinusoidal neuron’s influence across the domain. The result is sharper reconstructions, less noise, and faster convergence across image fitting, 3D volumes, and SDF reconstruction — consistently outperforming INR baselines.
Jun 1, 2026
Geometry-Guided Camera Motion Understanding in VideoLLMs
Accepted at CVPR 2026 Workshop PVUW. This work studies how to make VideoLLMs more camera-aware by benchmarking camera motion understanding and injecting geometry-derived motion cues at inference time.
Mar 13, 2026
Structured Pruning in Implicit Neural Representations
We are actively working on the formating of this paper.
Sep 15, 2024
Critical Features Tracking on Triangulated Irregular Networks by a Scale-Space Method
You can find the slides and a 15 mins presentation on the project page .
Sep 14, 2024
ImplicitTerrain: a Continuous Surface Model for Terrain Data Analysis
Please find the project page at .
Jun 18, 2024
Learning Discriminative Feature with CRF for Unsupervised Video Object Segmentation
Jan 1, 2020