Learning In-between Imagery Dynamics via Physical Latent Spaces
This work addresses the challenge of capturing evolving patterns in image data for applications like geoscience, though it appears incremental as it builds on existing latent space and PDE methods.
The paper tackles the problem of learning the underlying dynamics between two images at consecutive time steps by estimating intermediary stages, using a latent variable following a PDE-based physical model to ensure interpretability and preserve spatial correlations, with results demonstrated through numerical tests on geoscientific imagery data.
We present a framework designed to learn the underlying dynamics between two images observed at consecutive time steps. The complex nature of image data and the lack of temporal information pose significant challenges in capturing the unique evolving patterns. Our proposed method focuses on estimating the intermediary stages of image evolution, allowing for interpretability through latent dynamics while preserving spatial correlations with the image. By incorporating a latent variable that follows a physical model expressed in partial differential equations (PDEs), our approach ensures the interpretability of the learned model and provides insight into corresponding image dynamics. We demonstrate the robustness and effectiveness of our learning framework through a series of numerical tests using geoscientific imagery data.