Computational Imaging for Miniature Robots
|Interdisciplinary Areas:||Data and Engineering Applications, Engineering-Medicine, Autonomous and Connected Systems, Innovation and Making, CISLunar (Space science and Engineering), Defense related projects (for US citizens only), Smart City, Infrastructure, Transportation, Future Manufacturing, Micro-, Nano-, and Quantum Engineering, Power, Energy, and the Environment, Human-Machine/Computer Interaction, Human Factors, Human-Centered Design, Integrated Neuroscience and Engineering
People have been able to build small and low-power robots that mimic the mechanical behaviors of insects, such as bees , which could have broad applications in national security, agriculture, medicine, etc. However, it is still challenging to enable vision on these miniature platforms, as there have been limited visual sensors that are small and power-efficient enough to fit on these size-and-power-constrained platforms. This project will crack this conundrum by inventing new technologies, by integrating sensors, optics, and algorithms, that further reduce the form factors and power consumption of current visual sensors. The research will comprise three coupled aspects:
Computational optics, where we co-design optics with deep learning.
Camera image signal processing (ISP), and its integration with existing CMOS and photon counting image sensors.
Computer vision tools, such as NeRF and diffusion models.
The postdoc fellow will work with Prof. Stanley Chan and Prof. Qi Guo from the Department of Statistics and School of ECE, and a team of PhD students specializing in optical design and fabrication, and machine vision. The postdoc fellow will also collaborate with a broad range of faculties in the signal processing, machine learning, and optics areas of the School of ECE. The postdoc fellow will also have access to the cleanroom facilities at Purdue.
 Wood, R., Nagpal, R. and Wei, G.Y., 2013. Flight of the robobees. Scientific American, 308(3), pp.60-65.
January 1, 2024
PhD in Electrical Engineering, Computer Science, or related fields
Strong background in computational photography.
Strong knowledge in applied optics. Can conduct optical experiments independently.
Strong knowledge in building and utilizing machine learning models.
Excellent communication (oral and writing) skills.
Stanley Chan, Qi Guo
 Hazineh, D.S., Lim, S.W.D., Shi, Z., Capasso, F., Zickler, T. and Guo, Q., 2022. D-Flat: A Differentiable Flat-Optics Framework for End-to-End Metasurface Visual Sensor Design. arXiv preprint arXiv:2207.14780.
 Guo, Q., Shi, Z., Huang, Y.W., Alexander, E., Qiu, C.W., Capasso, F. and Zickler, T., 2019. Compact single-shot metalens depth sensors inspired by eyes of jumping spiders. Proceedings of the National Academy of Sciences, 116(46), pp.22959-22965.
 Stanley H. Chan, ‘‘What Does a One-Bit Quanta Image Sensor Offer?’’, IEEE Trans. Computational Imaging, vol. 8, pp. 770-783, Aug. 2022.
 Yiheng Chi, Xingguang Zhang, and Stanley H. Chan, ‘‘HDR Imaging with Spatially Varying Signal-to-Noise Ratios’’, IEEE Conf. Computer Vision Pattern Recognition (CVPR), 2023.
 Yash Sanghvi, Zhiyuan Mao, and Stanley H. Chan, ‘‘Structured Kernel Estimation for Photon-Limited Deconvolution’’, IEEE Conf. Computer Vision Pattern Recognition (CVPR), 2023.