Ep. 111 - January 29, 2024

arXiv Computer Vision research summaries for January 29, 2024. Today's Research Themes (AI-Generated): • A framework for downscaling tidal current data shows significantly improved predictions and reduced computational cost. • Large-scale bilingual image-text foundation models, M^2-Encoders, with increased efficiency and multi-language support for AI. • Multi-view video masked autoencoders introduce cross-view decoders to robustly handle motion and viewpoint changes. • CENet, a concise network, achieves a balance between accuracy and efficiency in image-guided depth completion for autonomous driving. • Novelty in flow image super-resolution achieved by incorporating quaternion spatial modeling and dynamic flow convolution.