M5-M Low Resolution and Quality Image Understanding

PI: Zhu LI

The goals of this project are (1) to make new end to end image processing pipeline that performs the extremely low light image denoising and enhancement task and (2) to develop a recognition friendly super resolution method for low resolution image recognition.

The project will utilize the following methodology:

  • Directly processing sensor field data
  • Learning based amplification of the dark
  • New residual learning based network for
    image denoising
  • Recovering super resolved gradient
    images at multiple scales to recover
    more information useful for high level
    vision tasks
  • CNN based super resolution networks
    to predict up scaled difference of
    Gaussian DoG images
  • Adapting the trained models to drive a
    widely adopted key point algorithm for
    image recognition