reconstruction accuracy
Reconstruction accuracy measures how well a model can reproduce input data from a compressed representation. This is a key metric in tasks involving autoencoders and generative models, where the goal is to capture essential features while minimizing distortion.
- DCI: Dual-Conditional Inversion for Boosting Diffusion-Based Image Editing
- Flexible MOF Generation with Torsion-Aware Flow Matching
- Generating Full-field Evolution of Physical Dynamics from Irregular Sparse Observations
- NPN: Non-Linear Projections of the Null-Space for Imaging Inverse Problems
- Test3R: Learning to Reconstruct 3D at Test Time
- Towards foundational LiDAR world models with efficient latent flow matching