CIF: Small: Signal Processing for Quanta Image Sensors: Reconstruction, Sampling, and Applications

CIF：小：Quanta 图像传感器的信号处理：重建、采样和应用

• 批准号: 1718007
• 负责人: Stanley Chan
• 金额: $ 48.09万
• 依托单位: Purdue University
• 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: 2017-07-01 至 2021-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1718007&HistoricalAwards=false
• 关键词: CIF; Small; Signal; Processing; Quanta

The proliferation of digital cameras over the past decades has revolutionized almost every aspect of science, technology and society. However, as we continue to shrink image sensors to increase spatial resolution and reduce camera size, the size of the photon-collecting site of a sensor will soon reach a fundamental limit below which no meaningful signal can be generated. The goal of this research is to investigate a new type of image sensor called the Quanta Image Sensors (QIS). Compared to conventional image sensors, QIS offer substantially higher spatial resolution, temporal resolution, and dynamic range. Because of these important features, the sensor has the potential to impact a wide range of applications in digital cameras, consumer electronics, wearable devices, medical imaging, and physical science. The focus of the research is to develop signal processing theories and algorithms to support QIS. Building upon previous studies in QIS, this research puts extra emphasis on bridging the theories to practice. Specific questions addressed include: (1) Image reconstruction. How to reconstruct images from the massive binary bit-stream collected by the QIS without using iterative algorithms? (2) Sampling. How to control the sensor?s sampling mechanism in spatial-temporal resolution, color space, and threshold to maximize the signal to noise ratio? (3) Application. How to track fast moving objects using QIS without reconstructing images? Results from this research will foster new signal processing techniques for imaging in photon limited environments.

在过去的几十年里，数码相机的普及几乎彻底改变了科学、技术和社会的各个方面。然而，随着我们不断缩小图像传感器以提高空间分辨率和减小相机尺寸，传感器的光子收集点的尺寸很快就会达到一个基本极限，低于这个极限就无法产生有意义的信号。本研究的目的是研究一种称为量子图像传感器（QIS）的新型图像传感器。与传统图像传感器相比，QIS提供了更高的空间分辨率、时间分辨率和动态范围。由于这些重要的功能，传感器有潜力影响在数码相机，消费电子产品，可穿戴设备，医学成像和物理科学的广泛应用。研究的重点是发展支持QIS的信号处理理论和算法。在以往QIS研究的基础上，本研究更加强调理论与实践的联系。具体解决的问题包括：(1)图像重建。如何在不使用迭代算法的情况下，从QIS收集的大量二进制比特流中重建图像？(2)抽样。如何控制传感器？S采样机制在时空分辨率，色彩空间，和阈值最大化信噪比？(3)应用程序。如何使用QIS跟踪快速移动的物体而不重建图像？这项研究的结果将促进光子有限环境下成像的新信号处理技术。

项目成果

Optimal Combination of Image Denoisers

• DOI: 10.1109/tip.2019.2903321
• 发表时间: 2019-08-01
• 期刊: IEEE TRANSACTIONS ON IMAGE PROCESSING
• 影响因子: 10.6
• 作者: Choi, Joon Hee;Elgendy, Omar A.;Chan, Stanley H.
• 通讯作者: Chan, Stanley H.

Image Reconstruction for Quanta Image Sensors Using Deep Neural Networks

• DOI: 10.1109/icassp.2018.8461685
• 发表时间: 2018-04
• 期刊: 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
• 作者: J. H. Choi;Omar A. Elgendy;Stanley H. Chan

Fast And Robust Recursive Filter for Image Denoising

• DOI: 10.1109/icassp.2018.8461887
• 发表时间: 2018-04
• 期刊: 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
• 作者: Yiheng Chi;Stanley H. Chan

Parameter-free Plug-and-Play ADMM for image restoration

• DOI: 10.1109/icassp.2017.7952371
• 发表时间: 2017-03
• 期刊: 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
• 作者: Xiran Wang;Stanley H. Chan

High dynamic range imaging with quanta image sensor

使用量子图像传感器进行高动态范围成像

• 发表时间: 2020
• 期刊: International Image Sensor Workshop (IISW
• 作者: Gnanasambandam, Abhiram;Ma, Jiaju;Chan, Stanley H
• 通讯作者: Chan, Stanley H