Eccv 2026 Fd2

Summary

TextDS is a parameter-efficient framework for robust scene text detection under domain shifts and adverse imaging conditions. Explainability-Aware Frustum Attack: Exposing Structural Vulnerabilities in LiDAR-Based 3D Object Detectors. Video for paper: VLA Knows Its Limits. Project link: hatchetproject.github.io/autohorizon/ Diffusion to Obfuscation: Time-Adaptive Synthesized Generation Against Gradient Leakage Attacks in Federated Learning. This is a presentation video of the paper: "FMA-Net++: Motion- and Exposure-Aware Joint Video Super-Resolution and ... NanoVSR: Towards Real-Time Video Super-Resolution on Edge Devices Paper: arxiv.org/abs/2607.10495 Code: ... FrozenDrive: Zero-Shot Text-Guided Driving Scene Generation and Data Augmentation with Parameter-Free Frozen Diffusion ...