lijenchang / nthu-computational-photography

NTHU EE6620 Computational Photography (Spring 2022)

This repository contains my code for the 3 assignments in EE6620 Computational Photography,
taught by Prof. Chao-Tsung Huang at National Tsing Hua University.

HW1 High Dynamic Range Imaging

Modern cameras are unable to capture the full dynamic range of commonly encountered natural scenes. High-dynamic-range (HDR) photographs are generally achieved by capturing multiple standard-exposure images, often using exposure bracketing, and then merging them into a single HDR image. Also, to view the HDR image on an ordinary low-dynamic-range (LDR) display, tone mapping operation from HDR to LDR on images is required. In this assignment, we implement the whole HDR photography flow, including image bracketing, camera response calibration [1], white balance, and finally, tone mapping, to visualize our results. More details can be found in the file hw1/EE6620-hw1.pdf.

HW2 Non-Blind Deblurring

A slow shutter speed will introduce blurred images due to camera shake. The objective of this assignment is to implement several non-blind deblurring algorithms and analyze the effects on blurred images. In part 1-3, we implement Wiener deconvolution, Richarson-Lucy (RL) deconvolution [2, 3] and its bilateral variant (BRL) [4]. And in part 4, we solve the deblurring problem by total variation regularization using ProxImaL [5]. More details can be found in the file hw2/EE6620-hw2.pdf.

HW3 Super-Resolution

Super-resolution is a class of techniques which can enhance the resolution of images. In this assignment, we work on two types of super-resolution methods, optimization-based and convnet-based. For the optimization-based part, we solve the SR problem with a single-image method and multi-image method using ProxImaL [5]. For the convnet-based part, we build and train a SR model based on WDSR [6], and then try to increase its size. More details can be found in the file hw3/EE6620-hw3.pdf.

References

1. P. E. Debevec and J. Malik. Recovering High Dynamic Range Radiance Maps from Photographs. In ACM SIGGRAPH, 2008.
2. W. H. Richardson. Bayesian-based Iterative Method of Image Restoration. In Journal of the Optical Society of America, 1972.
3. L. B. Lucy. An Iterative Technique for the Rectification of Observed Distributions. In Astronomical Journal, 1974.
4. L. Yuan et al. Progressive Inter-scale and Intra-scale Non-blind Image Deconvolution. In ACM Transactions on Graphics, 2008.
5. F. Heide et al. ProxImaL: Efficient Image Optimization Using Proximal Algorithms. In ACM Transactions on Graphics, 2016.
6. J. Yu et al. Wide Activation for Efficient and Accurate Image Super-Resolution. In arXiv preprint arXiv:1808.08718, 2018.