Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces GeoMag, a video motion magnification framework using State Space Models to achieve globally consistent motion amplification with linear complexity. To address the limitations of existing training data, the authors created Geo-200K, a large-scale synthetic dataset with rich geometric transformations and sensor-realistic degradations. Experiments demonstrate that GeoMag outperforms prior methods in visual fidelity, computational efficiency, and structural consistency on both synthetic and real-world benchmarks.
State Space Models can unlock globally consistent video motion magnification with linear complexity, outperforming CNNs and Transformers.
Video Motion Magnification (VMM) reveals imperceptible dynamics but often suffers from structural inconsistencies under complex geometric transformations. Existing learning-based methods generally face a trade-off between the limited global context of CNNs and the high computational cost of Transformers. In addition, current training protocols, largely dominated by simple linear motion, fail to capture the geometric and imaging complexities encountered in real-world videos. To address these issues, we propose GeoMag, a geometric-aware VMM framework built upon State Space Models to achieve globally consistent motion amplification with linear complexity. We further construct Geo-200K, a large-scale synthetic dataset that introduces rich geometric transformations together with sensor-realistic degradations, improving the diversity and realism of training signals. Extensive experiments on synthetic and real-world benchmarks show that GeoMag consistently outperforms prior methods in visual fidelity and computational efficiency, while producing fewer artifacts and better structural consistency.