AIxiv专栏是机器之心发布学术、技术内容的栏目。过去数年,机器之心AIxiv专栏接收报道了2000多篇内容,覆盖全球各大高校与企业的顶级实验室,有效促进了学术交流与传播。如果您有优秀的工作想要分享,欢迎投稿或者联系报道。投稿邮箱:liyazhou@jiqizhixin.com;zhaoyunfeng@jiqizhixin.com
论文标题:A Survey on Graph Neural Networks and Graph Transformers in Computer Vision: A Task-Oriented Perspective
论文地址:
https://arxiv.org/abs/2209.13232(预印版)
https://ieeexplore.ieee.org/document/10638815(IEEE 版)
建立在自然图像(二维)上的视觉任务包括 Image Classification (multi-label、few-shot、zero-shot、transfer learning),Object Detection,Semantic Segmentation,和 Scene Graph Generation。 建立在视频上的视觉任务包括 Video Action Recognition,Temporal Action Localization,Multi-Object Tracking,Human Motion Prediction,和 Trajectory Prediction。 视觉 + 语言方向的任务包括 Visual Question Answering,Visual Grounding,Image Captioning,Image-Text Matching,和 Vision-Language Navigation。 建立在三维数据上的视觉任务包括 3D Representation Learning (Point Clouds、Meshes),3D Understanding (Point Cloud Segmentation、3D Object Detection、3D Visual Grounding),和 3D Generation (Point Cloud Completion、3D Data Denoising、3D Reconstruction)。 建立在医学影像上的任务包括 Brain Activity Investigation,Disease Diagnosis (Brain Diseases、Chest Diseases),Anatomy Segmentation (Brain Surfaces、Vessels、etc)。