CSE 332 - Introduction to Visualization
Home
Schedule
Labs
Links
Policies
Grades

Labs

Slides from recitation session (5/13/09):
pdf files are available here:  isoSurface,   fullVolumeRendering

Lab 6 demo and more datasets (4/28/09):
Here is the demo and here are more datasets (carp, tooth, tree)

A few hints for lab5 (4/27/09):
Transferfunction: This time the transfer function serves as a lookup table / mapping function for the interpolated densities, d. They map the densities to RGBA.  Assume you would do nearest neighbor interpolation, you get:
sample.r=transfer_function[rint(d)].r, same for g,b,a. On the GPU the transferfunction is a 1D texture (it has 4 color channels), so you can interpolate them like any other texture, with tex1D(), and you use d as the index.

Ray step size and dimensions for volume texture maps:  For the ray step size, it depends what you choose for your texture extents. Choose this to create your 3D texture:
glTexImage3D(GL_TEXTURE_3D_EXT, 0, GL_RGBA, vol_w, vol_h, vol_d, 0, GL_RGBA, GL_UNSIGNED_BYTE, vox);
where vol_w, vol_h, vol_d are the volume extents. Now your ray step size will be intuitive. Typically you want to sample at least once per voxel cell, so you would choose step_size=1.0 or less.

X-ray min/max sliders:  Since X-ray sums densities, you can easily accumulate > 255 (just imagine your ray goes through a very dense area with many voxels having values of > 200, to give ray sums >> 255. Likewise you may also have rays that go through areas of low density, so you get raysums < 255. In the end you have a range of ray sums [0. 10000] or something like this, and this will get mapped to [0,255] at the display. You will not see good contrast. The windowing [min, max] will give you control over the density range where you will see good contrast, the rest will be saturated to 0 (rays<min) and 255 (rays>max).

Lab 5 demo, some more insight, and hig resolution volumes (4/13/09)
Here is a demo program that showcases MIP and X-ray rendering, but not the alpha-compositing (The option DVR is shaded volume rendering, to be modified for a future lab 6 demo).
Here are some more data files, they are hi-res versions of what you already have and will give you nicer images (but will take longer to render, which may not matter for a GPU implementation on a good machine).
Finally, here an updated lab5 GPU instructions file (the update is a section added to the beginning)

Instructions for GPU implementation of lab5 (4/8/09):
Here is a file that should make things quite clear.

Sample midterm from 2004 (4/6):
Here is a midterm exam given in 2004 which was the last tiomne this course was taught for undergrad students. It will closely resemble the style of the final exam. Please note that the final exam will and has always been comprehensive in scope.

FBO example (3/28/09):
Your class mate Zhiyuan Zhang was so kind and created a simple example application (available here) that shows how you write to an FBO and then copy the result to the framebuffer for display. Please note that the example uses unsigned byte as the data type. The FBO supports floats as well (but the display framebuffer does not), which you want to use for computations using the GPU. So you just change the FBO data type in this case.

Some  links to helpful CG information on the web (3/21):
The official NVIDIA CG tutorial (to get started -- advanced concepts such as FBOs are not discussed)
The NVIDIA developer site has many GL/CG examples
Some other links: here andhere
The CG Wiki page has also helpful links
Finally, here is a working program that uses keyboard commands to control the output (have a look at the source code)

Lab 3 (3/16/09)
Here is the added code, and here are the volume data files.

Lab 3 (2/26/09)
Here is the demo. It uses the graphics hardware. Make sure you have all libraries and files in the executable directory.

Lab 2 notes (2/24/09)
1. Modifying the histogram is a secondary effect. You modify the image with  the transfer function and this modifies the histogram (not the other way around). If the transfer function TF has for bin 23 (x=23) a value 50 (y=50) then this means that all image pixels with current value = 23 now will get value = 50. You need to make this change. The TF is an array [0..255] and has values [0...255]. Index it with your current image value and you will get out the new value.
2. Tthere is no direct relationship between transfer function and the histogram. In your program, they are only drawn into the same window and share the x-axis (0..255) since the x-axis holds the possible pixel values the input image can have. The y-axis for the transfer function is 0..255.

FLTK help (2/23/09)
See here for as file that shows how to add toggle and radio buttons and the functionality associated with it.

Lab 2 (2/19/09)
The demo is here. This demo uses the graphics hardware. Make sure you have all libraries and files in the executable directory.

Lab1
The skeletion code is here. Build with Visual Studio and you will get the basic functionality.
A FLTK user interface file, gui.fl, is also included to be used with the FLTK GUI builder package, fluid.
Finally, the distribution also contains a demo of the desired functionality, lab1.exe
A collection of images, but you are bery much encouraged to try your favorite images, or images you find opn the web, via google images or Flickr, etc  (see further below for conversion information).

New to C/C++?
All labs will use the C/C++ programming language.No worries, C/C++ is quite similar to Java. Here is a Tutorial on C/C++ for Java programmers

Need MS Visual Studio?
Go this website: https://www.dreamspark.com to get Visual Studio 2008 online. Must validate via student ID.

 Information on FLTK:
The following is for installation on MS Windows (2000, XP, Vista)
The GUI builder, fluid, will be in the fluid  folder .
Check out http://www.fltk.org for useful information.
Be sure to also check out the FLTK cheat sheet page for a great variety of FLTK code snippets to implement useful GUI functions 

Setup Visual Studio for use with FLTK:
Go to Tools -> Options-> Projects and Solutions (double-click to expose submenu list) -> VC Directories
Show directories: Include files (add the directory where the \FL folder is located)
      for example: C:\...\fltk 1.1.7\fltk-1.1.7\
Show directories: Library files (add the directory \lib in your FLTK folder hierarchy)
      for example: C:\...\fltk 1.1.7\fltk-1.1.7\lib
Now you are ready to select Build Solution to compile the executable (function key F5 will immediately execute the program)

Image file format: PPM:
The provided source code only supports ppm files, No matter if greylevel or color image, these images will always have an RGB triple for each pixel. The file format is (ASCII header, binary data):
   P6
   #Created by  (this line may not be there)
   Nx Ny (width and height of the image)
   255
   Binary pixel data (RGB, RGB,… ) 3 bytes per pixel, from the image bottom left to top right

Converting jpg, bmp, etc images to the PPM format:
Download and use Irfanview, a powerful public domain software that can so much more: image display and editing, screen and window capture, and much much more.