Engage Warp!

 
The Original Image

For Homework 5, we designed a program that would take user input and perform simple transformations or "warps" on an image. 

Overview

This project was designed to explore techniques that can be used to enhance the appearance of image warping results including Bilinear Interpolation and Super-Sampling.

Super-Sampling Using Filt Files

 For this project, I used a novel method to perform super-sampling that calculates the number and weight of the subpixel points based on a user-specified .filt file similar to the ones used in Assignment 4, Convolution Filtering.  

The number of sub-pixel sampling points is then equal to kernelSize².

Sines Original

Sines with Warp Only

Supersampling and Billinear interpolation
Box Filter - 9 Sample Points

Supersampling and Billinear interpolation
Box Filter - 81 Sample Points

Supersampling and Billinear interpolation
Tent 3 Filter - 9 Sample Points

Supersampling and Billinear interpolation
Tent Filter - 25 Sample Points

Supersampling and Billinear interpolation
Bell Filter - 81 Sample Points

 

---

Other Fun Results

The ability to specify a custom .filt file can produce interesting results when the filt files are not originally intended to blur the image. 

Sobel Horizontal

Sobel Vertical

For Assignment 2, we paired up in teams and created High Dynamic Range (HDR) Photos. You can check out our final work here:

http://www-viz.tamu.edu/courses/viza654/09fall/homework/Hines/hw02/index.html

There will be a contest among the other members of our class to see who has the best photo.

Our final photos are also listed below this post.

What is a High Dynamic Range photo?

Currently, most commonly available image display and capture technology can only display a certain limited range of the light values in a scene at any given time. Values that go over this range are simply mapped to "white," and values under this range are mapped to "black."

An example of this would be person a sitting inside in front of a window on a sunny day. If you focused on the scene outside, the scene inside would be mostly dark. However, if you focused inside, the light from the window would be completely blown out.

These  limited luminance-range images don't normally look weird to us because our eyes work the same way. Muscles in our eyes expand and contract our iris to let in more or less light in order to  reveal the details of whatever part of an environment we're looking at.

Since this process is mostly unconsious, however, traditional photos cannot replicate the "full beauty" and detail of certain vistas that we might experience in person.

High Dynamic Range photos attempt to compensate for this by combining multiple exposures of the same scene into one image. Each of these exposures capture different ranges of luminance values so that the full light value range can be represented. While the values of the final HDR image are still out of the display range of our modern devices, we can then "tone map" the image and compress the brightness values so that the details our eyes would reveal by expanding and contracting are maintained.