EAGER: Leveraging Synthetic Data for Visual Reasoning and Representation Learning with Minimal Human Supervision

EAGER：在最少的人类监督下利用合成数据进行视觉推理和表示学习

• 批准号: 1748387
• 负责人: Yong Jae Lee
• 金额: $ 20万
• 依托单位: University of California-Davis
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-08-15 至 2020-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1748387&HistoricalAwards=false
• 关键词: EAGER; Leveraging; Synthetic; Data; Visual

This project investigates how synthetic data created using computer graphics can be used for developing algorithms that understand visual data. Synthetic data provides flexibility that is difficult to obtain with real-world imagery, and enables opportunities to explore problems that would be difficult to solve with real-world imagery alone. This project develops new algorithms for reasoning about object occlusions, and for self-supervised representation learning, in which useful image features are developed without the aid of human-annotated semantic labels. The project provides new algorithms that have the potential to benefit applications in autonomous systems and security. In addition to scientific impact, the project performs complementary educational and outreach activities that engage students in research and STEM.This research explores novel algorithms that learn from synthetic data for visual reasoning and representation learning. While the use of synthetic data has a long history in computer vision, it has mainly been used to complement natural image data to solve standard tasks. In contrast, this project uses synthetic data to make advances in relatively unexplored problems, in which ground-truth is difficult to obtain given real-world imagery. The project consists of three major thrusts, each of which exploits the fact that a user has full control of everything that happens in a synthetic dataset. In Thrust I, it investigates a novel approach to representation learning using synthetic data, and in Thrust II, it extends the algorithm to disentangle task-specific and general-purpose features. Finally, in Thrust III, it explores a novel approach for reasoning about object occlusions.

这个项目研究如何使用计算机图形创建的合成数据可以用于开发理解视觉数据的算法。合成数据提供了现实世界图像难以获得的灵活性，并使探索仅使用现实世界图像难以解决的问题成为可能。该项目开发了用于对象遮挡推理的新算法，以及用于自监督表示学习的新算法，其中在没有人工注释语义标签的帮助下开发有用的图像特征。该项目提供了新的算法，有可能在自主系统和安全应用中受益。除了科学影响外，该项目还开展了补充教育和推广活动，吸引学生参与研究和STEM。本研究探索了从视觉推理和表征学习的合成数据中学习的新算法。虽然在计算机视觉中使用合成数据已经有很长的历史，但它主要用于补充自然图像数据以解决标准任务。相比之下，该项目使用合成数据在相对未探索的问题上取得进展，在这些问题上，给定真实世界的图像很难获得地面真相。该项目包括三个主要重点，每个重点都利用了这样一个事实，即用户可以完全控制合成数据集中发生的一切。在第一篇文章中，作者研究了一种使用合成数据进行表征学习的新方法，在第二篇文章中，作者扩展了该算法，以区分特定任务和通用功能。最后，在推力III中，它探索了一种新的方法来推理物体遮挡。

项目成果

HPLFlowNet: Hierarchical Permutohedral Lattice FlowNet for Scene Flow Estimation on Large-Scale Point Clouds

• DOI: 10.1109/cvpr.2019.00337
• 发表时间: 2019-06
• 期刊: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
• 作者: Xiuye Gu;Yijie Wang;Chongruo Wu;Yong Jae Lee;Panqu Wang

SinGAN-GIF: Learning a Generative Video Model from a Single GIF

• DOI: 10.1109/wacv48630.2021.00135
• 发表时间: 2021-01
• 期刊: 2021 IEEE Winter Conference on Applications of Computer Vision (WACV)
• 作者: Rajat Arora;Yong Jae Lee

You Reap What You Sow: Using Videos to Generate High Precision Object Proposals for Weakly-Supervised Object Detection

• DOI: 10.1109/cvpr.2019.00964
• 发表时间: 2019-06
• 期刊: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
• 作者: Krishna Kumar Singh;Yong Jae Lee

Boxer: Preventing fraud by scanning credit cards

Boxer：通过扫描信用卡防止欺诈

• 发表时间: 2020
• 期刊: USENIX Security Symposium
• 作者: Abi Din, Zainul;Venugopalan, Hari;Park, Jaime;Li, Andy;Yin, Weisu;Mai, Haohui;Lee, Yong Jae;Liu, Steven;King, Samuel
• 通讯作者: King, Samuel

MixNMatch: Multifactor Disentanglement and Encoding for Conditional Image Generation

• DOI: 10.1109/cvpr42600.2020.00806
• 发表时间: 2019-11
• 期刊: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
• 作者: Yuheng Li;Krishna Kumar Singh;Utkarsh Ojha;Yong Jae Lee