Computer Vision - ECCV 2026

D-VLAM: Differential Vision and Language Mixing for Rehearsal Free Continual Learning.- Mixture of Specialized Vision Experts: Unlocking Complementary Visual Insights for Faithful MLLM Reasoning.- Imaginative Perception Tokens Enhance Spatial Reasoning in Multimodal Language Models.- PanoRec: Spatially-Structured Sequence Modeling for Multi-Granularity Panoramic Retrieval.- Residual-Guided Expert Specialization for Incomplete Multimodal Learning.- Flash-Refine: Frustum-Guided Local Incremental Learning for Efficient 3D Gaussian Splatting Completion.- OmniFall: From Staged Through Synthetic to Wild, A Unified Multi-Domain Dataset for Robust Fall Detection.- LiFlow: Flow Matching for 3D LiDAR Scene Completion.- Are Video Reasoning Models Ready to Go Outside?.- UC-VLM: Consistency-Driven Learning for AI-Generated Image Detection with Vision-Language Large Models.- Track4World: Feedforward World-centric Dense 3D Tracking of All Pixels.- ARC-Loc: Leveraging Azimuthal Ray Convergence as a Geometric Cue for Direct Cross-View Localization.- DAPS++: Rethinking Diffusion Inverse Problems with Decoupled Posterior Annealing.- EvoVLA: Self-Evolving Vision-Language-Action Model.- DP-BOA: Dirichlet-Process Birth-or-Assign for On-the-Fly Category Discovery.- Fair and Faithful: A Diffusion-Enhanced Dataset and Hybrid State-Space Mamba for Face Super-Resolution.- ReCamDriving: LiDAR-Free Camera-Controlled Video Synthesis for Novel Trajectories.- PolyLayout: Multi-room Manhattan Layout Estimation.- Better Call CineCrew: Consistent Ultra-Long Narrative-to-Film Generation.- Histogram-constrained Image Generation.- EraseLoRA: MLLM-Driven Foreground Exclusion and Background Subtype Aggregation for Dataset-Free Object Removal.- Uncertainty-aware tree height change regression.- DriveFine: Refining-Augmented Masked Diffusion VLA for Accurate and Robust Driving.- PhysEdit: Physically Consistent Image Editing via Causal Enforcement.- Spectral-Aware Analytic Class-Incremental Learning for Long-Tailed Distributions.- Why Can Accurate Models Be Learned from Inaccurate Annotations?.- JanusMesh: Fast and Zero-Shot 3D Visual Illusion Generation via Cross-Space Denoising.- Dual Distribution Estimation for Zero-shot Noisy Test-Time Adaptation with VLMs.- OCTOPUS: Multi‑Agentic Universal Compositional Visual Retrieval.- Diversity-Aware View Partitioning for Scalable VGGT.- WildCity: A Real-World Dataset for City-Scale Rendering and Beyond.- Lifting Ego World Models for Planning and Control.- G2FM: A Geodesic Flow Matching Framework with Geometric Prior for Category-Level 9-DoF Pose Estimation.- Personalizing MLLMs via Reinforced Multimodal Reference Game.- Rethinking IRSTD: Single-Point Supervision Guided Encoder-only Framework is Enough for Infrared Small Target Detection.- MAC-Splat: Multi-Attribute Consistency for High-Fidelity Sparse-View Reconstruction.- Learning from Primitive: Probing Visual Reasoning of LVLMs via Counting.

• Erscheinungsdatum: 10.09.2026
• Autor/Autorin: Paolo Favaro
• Format: E-Book
• Dateiformat: PDF
• Kopierschutz: Wasserzeichen
• Dateigröße: 165.2 MB
• Verlag: SPRINGER
• Sprache: Englisch
• Umfang: 702 Seiten
• ISBN: 9783032370235
• Lieferung: Sofort per Download
• Hinweis: Sofort per Download lieferbar. Kein physischer Versand.
• Kompatibilität: Lesbar auf Geräten und Apps mit PDF-Unterstützung.