全部 |
  • 全部
  • 题名
  • 作者
  • 机构
  • 关键词
  • NSTL主题词
  • 摘要
检索 二次检索 AI检索
外文文献 中文文献
筛选条件:

1. Enhancing Recipe Retrieval with Foundation Models: A Data Augmentation Perspective NSTL国家科技图书文献中心

Fangzhou Song |  Bin Zhu... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 111~127 - 共17页

摘要:Learning recipe and food image representation |  in common embedding space is non-trivial but |  crucial for cross-modal recipe retrieval. In this paper | , we propose a new perspective for this problem by |  utilizing foundation models for data augmentation
关键词: Recipe retrieval |  Data augmentation |  Foundation models

2. SCOMatch: Alleviating Overtrusting in Open-Set Semi-supervised Learning NSTL国家科技图书文献中心

Zerun Wang |  Liuyu Xiang... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 217~233 - 共17页

摘要:Open-set semi-supervised learning (OSSL | ) leverages practical open-set unlabeled data, comprising |  both in-distribution (ID) samples from seen classes |  and out-of-distribution (OOD) samples from unseen |  classes, for semi-supervised learning (SSL). Prior OSSL
关键词: Open-set problem |  Semi-supervised learning

3. Image Compression for Machine and Human Vision with Spatial-Frequency Adaptation NSTL国家科技图书文献中心

Han Li |  Shaohui Li... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 382~399 - 共18页

摘要:Image compression for machine and human vision |  (ICMH) has gained increasing attention in recent years | . Existing ICMH methods are limited by high training and |  storage overheads due to heavy design of task-specific |  networks. To address this issue, in this paper, we
关键词: Learned image compression |  Machine vision |  Adapter

4. Efficient Snapshot Spectral Imaging: Calibration-Free Parallel Structure with Aperture Diffraction Fusion NSTL国家科技图书文献中心

Tao Lv |  Lihao Hu... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 93~110 - 共18页

摘要:Aiming to address the repetitive need for |  complex calibration in existing snapshot spectral |  imaging methods, while better trading off system |  complexity and spectral reconstruction accuracy, we |  demonstrate a novel Parallel Coded Calibration-free Aperture
关键词: Calibration-free |  Aperture diffraction fusion |  Snapshot spectral imaging

5. Continuous Memory Representation for Anomaly Detection NSTL国家科技图书文献中心

Joo Chan Lee |  Taejune Kim... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 438~454 - 共17页

摘要:There have been significant advancements in |  anomaly detection in an unsupervised manner, where only |  normal images are available for training. Several |  recent methods aim to detect anomalies based on a |  memory, comparing or reconstructing the input with
关键词: Anomaly detection |  Continuous memory representation

6. Coarse-to-Fine Implicit Representation Learning for 3D Hand-Object Reconstruction from a Single RGB-D Image NSTL国家科技图书文献中心

Xingyu Liu |  Pengfei Ren... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 74~92 - 共19页

摘要:Recent research has explored implicit |  representations, such as signed distance function (SDF), for |  interacting hand-object reconstruction. SDF enables modeling |  hand-held objects with arbitrary topology and |  overcomes the resolution limitations of parametric models
关键词: 3D hand-object reconstruction |  RGB-D fusion |  Implicit representation

7. StructLDM: Structured Latent Diffusion for 3D Human Generation NSTL国家科技图书文献中心

Tao Hu |  Fangzhou Hong... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 363~381 - 共19页

摘要:Recent 3D human generative models have |  achieved remarkable progress by learning 3D-aware GANs |  from 2D images. However, existing 3D human generative |  methods model humans in a compact 1D latent space | , ignoring the articulated structure and semantics of human
关键词: 3D human generation |  Latent diffusion model

8. NeuroPictor: Refining fMRI-to-Image Reconstruction via Multi-individual Pretraining and Multi-level Modulation NSTL国家科技图书文献中心

Jingyang Huo |  Yikai Wang... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 56~73 - 共18页

摘要:Recent fMRI-to-image approaches mainly focused |  on associating fMRI signals with specific |  conditions of pre-trained diffusion models. These |  approaches, while producing high-quality images, capture |  only a limited aspect of the complex information in
关键词: Neural decoding |  FMRI-to-Image |  Diffusion model

9. On the Approximation Risk of Few-Shot Class-Incremental Learning NSTL国家科技图书文献中心

Xuan Wang |  Zhong Ji... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 162~178 - 共17页

摘要:Few-Shot Class-Incremental Learning (FSCIL | ) aims to learn new concepts with few training samples |  while preserving previously acquired knowledge | . Although promising performance has been achieved, there |  remains an underexplored aspect regarding the basic
关键词: Few-Shot class-Incremental learning |  Few-Shot learning |  Class incremental learning |  Approximation risk

10. General Geometry-Aware Weakly Supervised 3D Object Detection NSTL国家科技图书文献中心

Guowen Zhang |  Junsong Fan... -  《Computer Vision - ECCV 2024,Part LI》 -  European Conference on Computer Vision - 2025, - 290~309 - 共20页

摘要:3D object detection is an indispensable |  component for scene understanding. However, the annotation |  of large-scale 3D datasets requires significant |  human effort. To tackle this problem, many methods |  adopt weakly supervised 3D object detection that
关键词: Weakly supervised object detection |  3D object detection |  General scenarios
检索条件出处:Computer Vision - ECCV 2024,Part LI
  • 检索词扩展

NSTL主题词

  • NSTL学科导航