CS 766 Computer Vision Assignment 1 - Spring 2014

Project done by : Pradeep Rajendran, Gautam Umesh Sargur

Aim

This project aims to recover high dynamic range radiance maps from a series of bracketed photographs taken with different exposures and use these radiance maps to construct a High Dynamic range image and view it by tone mapping it.

Algorithms implemented

1. Recovering high dynamic range radiance maps(one for each channel of RGB) from series of digital images taken at different exposures using Debevec's algorithm. This is obtained by an intermediatery step from which we obtain the response curve of the camera used
2. Output the high dynamic range radiance maps as .hdr file and tonemap it to display the final HDR image.
3. Durand Algorithm for tonemapping the image. (bonus)
4. Using Exposure fusion technique to get better HDR images (bonus)

Image Set used to get HDR image

We took one photo set, but this was done with due care to ensure that we would require one. The image set contains 8 images taken with various exposures. The images basically contain two buildings, one under direct sunlight and the other in the shade resulting in details being lost due to over-exposure in the builing under direct sunlight and not enough detail being captured in the building in the shade because of under-exposure resulting in imbalance in the amount of light radiating from different surfaces in the image. The main objective was to retrieve as much details as possible from both structures in the image while not compromising the quality of the image itself.

Approach

• Getting the camera response curve and radiance maps: We got the camera response curve by imlementing Debevec's algorithm. Then we got the log irradiance value for each pixel in each image by mapping the pixels values using the response curve. Though ideally, the log irradiance should remain constant across exposures, we use a triangular weighting function for robustness. The log irradiance and by extension the irradiance values for each pixel will be a weighted average of pixels from all input images. All this is done in MATLAB. There are two versions of the code as both of us implemented it independantly.


• The RGBE formatted HDR file was output from MATLAB along with a tonemapped version using MATLAB's inbuilt tonemap() function.


• We also tonemapped the HDR image output from MATLAB using Reinhard's Algorithm in Luminance HDR. We show a comparative result of the difference in tonemapping we got from the MATLAB version versus the Reinhard Algorithm.


• (Bonus) We used an algorithm called exposure fusion as described in this paper to obtain a image which is the fusion of several exposures and re-insert this image into the HDR algorithm with average exposure time all input images and we show a comparative result of using an exposure fused image vs. not using it. This implementation of this algorithm was borrowed from the code made public by the authors of this paper. We however, show the weight maps of Saturation, Contrast and Weel-exposedness, which are the three measures used by the paper to fuse these images.


• (Bonus) We also implemented Durand algorithm as described in this paper as an alternative method of tonemapping to Reinhard's algorithm.


• Instructions on how to run the code and the functionality of all files has been explained in detail in rhe README file submitted along with the code.

Input Photo Set We took:

Response curve of the Canon Powershot A640 is like this:

Exposure Fusion:

Original Input Images
Saturation Weight Maps.
Contrast Weight Maps.
Well-Exposedness weight maps.

The Final tonemapped output and comparison:

Tonemapping using Reinhard's Algorithm.	Output of Exposure Fusion algorithm itself.	Reinhard's Algorithm including Exposure fused image.	Tonemapping in MATLAB.	Tonemapping Using Durand Algorithm.

Performance of our algorithm on other images (Source: http://www.easyhdr.com)

Individual Contributions

Pradeep Rajendran : Recovering HDR maps and tonemapping, Input image Acquisition, Durand Algorithm implementation

Gautam Umesh Sargur : Recovering HDR maps and tonemapping, Using the Exposure fusion algorithm in HDR, Making the HTML report

References

[1] Paul E. Debevec, Jitendra Malik, Recovering High Dynamic Range Radiance Maps from Photographs, SIGGRAPH 1997

[2] Tom Mertens, Jan Kautz, Frank Van Reeth1, Exposure Fusion

[3] Fredo Durand, Julie Dorsey, Fast Bilateral Filtering for the Display of High-Dynamic-Range Images