Schweitzer Fachinformationen
Wenn es um professionelles Wissen geht, ist Schweitzer Fachinformationen wegweisend. Kunden aus Recht und Beratung sowie Unternehmen, öffentliche Verwaltungen und Bibliotheken erhalten komplette Lösungen zum Beschaffen, Verwalten und Nutzen von digitalen und gedruckten Medien.
The multi-volume set of LNCS books with volume numbers 15059 up to 15147 constitutes the refereed proceedings of the 18th European Conference on Computer Vision, ECCV 2024, held in Milan, Italy, during September 29-October 4, 2024.
The 2387 papers presented in these proceedings were carefully reviewed and selected from a total of 8585 submissions. The papers deal with topics such as computer vision; machine learning; deep neural networks; reinforcement learning; object recognition; image classification; image processing; object detection; semantic segmentation; human pose estimation; 3d reconstruction; stereo vision; computational photography; neural networks; image coding; image reconstruction; object recognition; motion estimation.
Rethinking Data Bias: Dataset Copyright Protection via Embedding Class-wise Hidden Bias.- Pose-Aware Self-Supervised Learning with Viewpoint Trajectory Regularization.- SILC: Improving Vision Language Pretraining with Self-Distillation.- Learning Semantic Latent Directions for Accurate and Controllable Human Motion Prediction.- Leveraging temporal contextualization for video action recognition.- ChEX: Interactive Localization and Region Description in Chest X-rays.- AdaGlimpse: Active Visual Exploration with Arbitrary Glimpse Position and Scale.- CLAP: Isolating Content from Style through Contrastive Learning with Augmented Prompts.- ZigMa: A DiT-style Zigzag Mamba Diffusion Model.- EchoScene: Indoor Scene Generation via Information Echo over Scene Graph Diffusion.- On Calibration of Object Detectors: Pitfalls, Evaluation and Baselines.- HAT: History-Augmented Anchor Transformer for Online Temporal Action Localization.- Deep Nets with Subsampling Layers Unwittingly Discard Useful Activations at Test-Time.- Safe-Sim: Safety-Critical Closed-Loop Traffic Simulation with Diffusion-Controllable Adversaries.- Analysis-by-Synthesis Transformer for Single-View 3D Reconstruction.- Challenging Forgets: Unveiling the Worst-Case Forget Sets in Machine Unlearning.- WaSt-3D: Wasserstein-2 Distance for Scene-to-Scene Stylization on 3D Gaussians.- SCLIP: Rethinking Self-Attention for Dense Vision-Language Inference.- Flying with Photons: Rendering Novel Views of Propagating Light.- RGNet: A Unified Clip Retrieval and Grounding Network for Long Videos.- MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Images.- 3DGazeNet: Generalizing Gaze Estimation with Weak Supervision from Synthetic Views.- Removing Distributional Discrepancies in Captions Improves Image-Text Alignment.- Resilience of Entropy Model in Distributed Neural Networks.- Rejection Sampling IMLE: Designing Priors for Better Few-Shot Image Synthesis.- Implicit Concept Removal of Diffusion Models.- PLOT: Text-based Person Search with Part Slot Attention for Corresponding Part Discovery.
