EasyTrain instructions:

Usage:
  java EasyTrain [-isYUV | -isRGB] [-show YUV|HSB|rg|xy|Lab] imagefile...

See additional tutorials at:
  http://www.cs.cmu.edu/~dst/Tekkotsu/Tutorial/colorsegment.shtml
  http://www.cs.cmu.edu/~tekkotsu/CameraSetup.html


== Command Line Flags: ===========

  -isYUV, -isRGB   Report the mode that input images were saved in.

  -show <mode>     Controls the color space shown in Color Spectrum window:
                     YUV - displays U,V channels; native space of the AIBO
                     HSB - [default] displays Hue and Saturation
                     rg  - R and G contributions of RGB space
                     xy  - X and Y contributions of XYZ space
                     Lab - displays a,b channels

  Among the color spaces available for showing, all except HSB distribute
  hues basically radially, with a white point at the center.  HSB places
  the hues along the x axis, and saturation along the y; thus the white
  point is spread out along the top of the plot.
      YUV might be preferred for its direct control
      HSB might be preferred for a horizontal view of hue
      Lab might be preferred for a radial view of hue
      (rg and xy are mainly side-effects, but made available if desired)


== Expected Image Format: ========

  Image files should be saved from the Tekkotsu RawCam window as PNG files
  with YUV encoding, but the program can also handle images in RGB
  encoding.  If you do decide to use JPEG images, you should actually save
  them in 'RGB' mode, because the JPEG format uses YUV internally.

  Pass '-isYUV' if the images were saved in YUV mode, pass '-isRGB' if they
  were saved in RGB mode.


== Window List: ==================

1. RGB Image View: displays the current image file in RGB color space.  Use
   the arrow keys to cycle through all the image files.

2. Segmented Viewer: displays the color-segmented version of the
   current image file.  (If no colors are defined yet, it will be all
   grey.)

3. Color Spectrum Window: displays pixels from selected image regions in a
   2D color space representation.  Draw a bounding region with the mouse to
   select pixel clusters.  See discussion of the '-show' option for
   controlling the color space used

4. Control window: contains buttons to add or remove colors, load or save
   files, etc.  Hover over items for tooltips.

5. Thumbnail window: shows thumbnails of loaded images, allows you to
   directly jump to specific images by clicking on them.


== Shortcuts: ==============

  In any of the editor windows (1-3 above):
    Ctrl-A: select all
    Ctrl-D: deselect
    Ctrl-Z: undo
    Shift-Ctrl-Z: redo
    '+': Increase magnification multiple
    '-': Decrease magnification multiple
    '=': Reset magnification to "actual size"


== Usage: ==================
 
  Start by entering a color name, say, "pink".  (Type the color name in the
  dialog box at the top of the GUI window, and hit "enter".)

  Now use the mouse to circle a pink region in the raw image window.  This
  will cause a cloud of pink pixels to appear in the color space window.
  (Note: if you have a pink object on a green background and you just draw
  a big box around it, so that some green pixels are included in the box,
  that's okay, because the green pixels will be far enough away from the
  pink ones in color space that you can easily separate them.)

  Use the mouse to circle just the cloud of pink pixels in color space.
  Once you've done that, the pink regions should appear in the color
  segmented image window.

  You can add to a region by holding down the shift key and drawing with
  the mouse.  You can subtract from a region by holding down the control
  key and drawing.

  You can use the arrow keys in the raw or segmented image windows to cycle
  though the images.  You can circle zero or more regions in each image.
  Pixels from all the images will be lumped together in the color space
  window.  (If no regions are selected, or 'all pixels' control is
  selected, then the full color spectrum is shown)

  After you've defined all your colors, use the "Save" button to save your
  work.  The area boundary info for each image will be saved, along with
  the .col and .tm files needed by CMVision.  That means if you come back
  later and run the program again, loading in the same image files, the
  program will also load in the boundaries you defined, so you can continue
  to edit them.

================

Note to Linux users: if you're receiving warnings about inability to locate
a system preferences file, this is because of the way Java was installed.
See http://java.sun.com/j2se/1.5.0/install-linux.html for a fix.