CRII: RI: Learning with Low-Quality Visual Data: Handling Both Passive and Active Degradations

CRII：RI：使用低质量视觉数据学习：处理被动和主动退化

• 批准号: 1755701
• 负责人: Zhangyang Wang
• 金额: $ 17.3万
• 依托单位: Texas A&M Engineering Experiment Station
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-08-15 至 2020-10-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1755701&HistoricalAwards=false
• 关键词: CRII; RI; Learning; Low; Quality

This project is focused on effectively and robustly exploiting low-quality (LQ) visual data for computer vision tasks. While most current computer vision systems are designed for high-quality visual data, collected from "clear" environments where subjects are well observable without significant attenuation or alteration, a dependable vision system must reckon with the entire spectrum of degradations from unconstrained environments. With various degradations arising from the visual data acquisition and processing pipeline, the ubiquitous LQ visual data can dramatically deteriorate the model performance in practice. The project outcome can broadly benefit a variety of real-world applications, such as video surveillance, autonomous/assisted driving, robotics and medical image analysis, where LQ visual data has constituted major performance and reliability bottlenecks. This research categorizes common degradations into the two types: "passive degradations" that are caused by uncontrollable environment factors (such as bad weather and low light); and "active degradations" that are intentionally introduced in a controllable way to meet certain budget requirements (such as lossy compression). The project will mainly addresses two important technical questions: i) how to overcome passive degradations and achieve more robust high-level task performance on LQ video data, using end-to-end deep learning models; and ii) how to properly introduce and control active degradations to generate the desired form of LQ data, that both satisfies certain budget requirements and maintains the target task utility, using deep adversarial learning models. The resulting new techniques are to be verified on application examples such as video recognition, video annotation, video compression, and de-identified video data sharing for recognition purpose.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

该项目专注于有效、稳健地利用低质量 (LQ) 视觉数据来执行计算机视觉任务。虽然大多数当前的计算机视觉系统都是为高质量的视觉数据而设计的，这些数据是从“清晰”的环境中收集的，在这些环境中，可以很好地观察对象，而不会出现明显的衰减或改变，但可靠的视觉系统必须考虑到不受约束的环境造成的整个退化范围。由于视觉数据采集和处理管道产生各种退化，无处不在的 LQ 视觉数据会极大地降低实践中的模型性能。该项目成果可广泛惠及各种实际应用，例如视频监控、自动/辅助驾驶、机器人和医学图像分析，其中 LQ 视觉数据构成了主要的性能和可靠性瓶颈。本研究将常见的退化分为两类：由不可控的环境因素（如恶劣天气和弱光）引起的“被动退化”；以及以可控方式有意引入的“主动降级”以满足某些预算要求（例如有损压缩）。该项目将主要解决两个重要的技术问题：i）如何使用端到端深度学习模型克服被动退化并在 LQ 视频数据上实现更稳健的高级任务性能； ii）如何使用深度对抗学习模型正确引入和控制主动降级以生成所需形式的 LQ 数据，既满足某些预算要求又保持目标任务效用。由此产生的新技术将在视频识别、视频注释、视频压缩和用于识别目的的去识别化视频数据共享等应用实例上进行验证。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准。

项目成果

DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Better

DeblurGAN-v2：更快更好的去模糊（数量级）

• DOI: 10.1109/iccv.2019.00897
• 发表时间: 2019-08-10
• 期刊: 2019 IEEE/CVF International Conference on Computer Vision (ICCV)
• 作者: Orest Kupyn;T. Martyniuk;Junru Wu;Zhangyang Wang
• 通讯作者: Zhangyang Wang

Single Image Deraining: A Comprehensive Benchmark Analysis

单幅图像去雨：综合基准分析

• DOI: 10.1109/cvpr.2019.00396
• 发表时间: 2019-03-20
• 期刊: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
• 作者: Siyuan Li;I. B. Araujo;Wenqi Ren;Zhangyang Wang;Eric K. Tokuda;Roberto Hirata Junior;R. M. C. Junior
• 通讯作者: R. M. C. Junior

Towards Privacy-Preserving Visual Recognition via Adversarial Training: A Pilot Study

通过对抗性训练实现保护隐私的视觉识别：一项试点研究

• DOI: 10.1007/978-3-030-01270-0_37
• 发表时间: 2018-07-22
• 作者: Zhenyu Wu;Zhangyang Wang;Zhaowen Wang;Hailin Jin
• 通讯作者: Hailin Jin

Enhance Visual Recognition under Adverse Conditions via Deep Networks

通过深度网络增强不利条件下的视觉识别

• DOI: 10.1109/tip.2019.2908802
• 发表时间: 2019-01
• 期刊: IEEE Transactions on Image Processing
• 影响因子: 10.6
• 作者: Liu, Ding;Cheng, Bowen;Wang, Zhangyang;Zhang, Haichao;Huang, Thomas S.
• 通讯作者: Huang, Thomas S.