Temporal Periodic Image Registration with Implicit Neural Representations
In this work, we extend INRs for deformable image registration to the domain of temporal image registration, focusing on periodic temporal image sequences.
ProofPaper ↗
Key points
- Implicit Neural Representations (INRs) have recently gained popularity for their ability to model functions using simple and lightweight networks.
- These methods have also proven effective in deformable image registration.
- Our approach, Temporal-IDIR, optimizes a single INR to model deformations across all frames in a temporal image sequence simultaneously, allowing for self-regularization through its own deformation predictions.
- To achieve this, we introduce a temporal consistency loss that penalizes discrepancies between direct source-to-target transformations and those traversing intermediate frames.
Sources (1)
- [1]Temporal Periodic Image Registration with Implicit Neural RepresentationsarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 01:24 PM
In this work, we extend INRs for deformable image registration to the domain of temporal image registration, focusing on periodic temporal image sequences.
Implicit Neural Representations (INRs) have recently gained popularity for their ability to model functions using simple and lightweight networks.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 15, 2026FuseAdapt: Parameter-Efficient Multimodal Fusion for Semantic Segmentation with Missing Modalities
- Oct 15, 2026StochLOB: Heavy-Tailed Score Matching for Robust Limit Order Book Trend Prediction
- Oct 15, 2026Event-STVG: Spatio-Temporal Video Grounding for Long-Form Event Reasoning
- Oct 15, 2026Learning to Think and Explore in 3D World
- Oct 14, 2026mistralai/Timestral-209M
- Oct 13, 2026Correction Training for Diffusion Models