Contents || Index
03/20/2007

 

Tutorial Contents | Tutorial 1 | Tutorial 2

Tutorial 3: Advanced Imaging and Spectral Analysis

The basic production of sum spectra, images and color overlays encompasses most of the functions required by common Lispix users.  However, Lispix contains numerous tools for further investigating hyperspectral data.  These tools include the separation of phases and the production of single bit (black and white) images known as “masks” for the identification of compounds in the system.

This tutorial will show the user how to:

The Lispix Window


Rectangles and User Defined Shapes

Phases that lie in distinct regions in an X-ray map are frequently of interest to more advanced users.  Lispix has tools to select groups of pixels based on user defined criteria.  The simplest tool is the rectangle.  Click the image to activate it, then in the “Select” menu, click “Draw Rectangle”.  Clicking and dragging on an image will produce a yellow box on the image. 

After drawing, the rectangle can be moved by clicking within the rectangle and dragging.

In addition, the same rectangle can be shown on multiple images.  Click on an image to activate it, and then click Select / Show Rectangle.  A new rectangle will appear in the same place on the newly activated image.  The new rectangle can also be moved (but not resized) by clicking Select / Adjust Rectangle, and then clicking and dragging within the rectangle. 

Rectangles can be forced to squares, set by typing and selected as a predefined size by using the tools in the “Select” menu on the main toolbar.

Once a rectangle is drawn, Lispix can also crop the image to that rectangle.  This is useful for cropping large images or highlighting certain sections.  Click Select / Crop to Rectangle to crop the image.

 

Generating spectra from rectangles

Once the correct rectangle is drawn, the Data Cube tool has functions that will derive sum spectra from pixels within the rectangle.  Selecting Tools / Data Cube will open the tool.  Click Plots / Sum Spect Within Rectangle to derive a sum spectrum from pixels in the rectangle.

 

Information about the rectangle

In addition to generating information about pixels inside the rectangle, Lispix will also produce statistics and information about the rectangle.  This is useful when counting pixels in an image (to determine size with length standards) or when sizing features in the image.  If k-ratio or concentration maps are created (see tutorial 4), the rectangle information will give valuable information such as average value, standard deviation and max/min.

Once the desired rectangle is created, click Info / Rect Stats to bring up the rectangle statistics window.  If the rectangle is adjusted using the Select / Adjust command, the Info / Rect Stats will bring up a new text window to display the rectangle statistics.

With all text windows, the text can be copied to a word processing program or saved as an ASCII text file using the File / Save Text command.  If used, the save text command will save the text of the active window.  Click anywhere on the text window to activate it.

 

 

Single pixel information

Lispix has tools to allow the user to select single pixels in an image.  For all of the commands listed in this section, remember to click an image to activate it.  Use the command Select / Click Single Pixel to choose a single pixel.  Lispix will draw a yellow rectangle around the pixel of choice.  The command Info / Rect Stats will give the statistics and information for the single highlighted pixel.

If the user does not wish to highlight pixels, Lispix has three tools to allow the user to gain pixel information instantly by clicking on them. 

The command Info / Pixel Info on Click will open a text dialogue box displaying the x and y position of the pixel as well as the value.  In most images, this value is the brightness of the pixel.  Clicking on numerous pixels in the image will generate a list of pixels.  It should be noted that if the user clicks on the same pixel twice, Lispix will report that pixel value twice without notifying the user.
The command Info / Statistics on Click will first open a dialogue box asking the user to determine the pixel radius to base the statistics on.  The radius is a value of pixels away from the clicked pixel.  Lispix creates a square around the click and includes all pixels within the square.  Clicking repeatedly on the image will generate multiple circles each with a separate statistics group.  Like the “Pixel Info on Click” command, the “Statistics on Click” command will not warn the user of overlapping pixels.  Additionally, no markings will appear on the image to show what pixels have been analyzed or the extent of the pixel square.
The command Info / Chart on Click will produce a chart of all values around a pixel, rather than an average or statistical grouping of the pixels.  The command will first open a dialogue box asking for a radius of pixels around which to draw the square.  The next dialogue box will ask the user to select a format for the numbers.  For instance, the term “Fixed 4” will produce integer values with 4 spaces between each column.  The author suggests “Fixed 4” as a simple, readable output for this function.  The float options will produce values to 1 or 2 decimals places.

 

Generating Masks

One topic not covered in Tutorial 2 was the use of the “Threshold Slider” found in the “Colorize” menu.  This slider is used to separate groups of pixels based on brightness.  The user selects a range of values using the slider, and the function colors pixels in the range red. 

Open the slider by choosing Colorize / Threshold Slider on the menu bar.  The slider will appear at the bottom of the active image.  When clicked, the pixels in range (the value range is displayed on the right side of the image) will colorize.  Like the region slider, the limits and position of the slider may be changed by dragging on the edges and in the center.

When the correct area has been highlighted, click the “thresholds” button and select “make mask”.  A binary image will appear with “mask” preceding the title of the window.

The constituents of the mask may be further processed to determine the analytical composition of the mask.  In the “Data Cube” tool, click Plot / Sum Spect with Mask to create a sum spectrum of all the pixels in the mask.  A dialogue box will appear asking the user to select which image is the desired mask.  Refer to the top of the image window to determine the name of the desired image.

 

 

 

 

Background Subtracted Images

Images of minor elements or compounds may appear grainy due to issues with auto-scaling.  Lispix has tools to eliminate background data from the image, providing a more accurate representation of the image.

In the “Data Cube” tool, click the “Sliders” button and select “Peak – Background”.  This will open up a new slider tool below the active plot. 

The background sliders are colored in gray, and the peak slider is colored in pink.  Move the background slider to areas where the background is linear and representative of the actual background below the peak.  Lispix will average the values in both sliders to linearly interpolate the background below the peak. 

Once the appropriate areas are set up, click the “Image” button in the “Data Cube” tool and select “Peak Minus Background”.  Lispix will then generate the background subtracted image.

 

 

Cheat Sheet

  1. Draw Rectangles
    1. Select / Draw Rectangle
    2. Click and drag to form yellow rectangle on the image
    3. Select / Adjust Rectangle
    4. Click rectangle and drag to move rectangle over image
    5. Click image to activate
    6. Select / Show Rectangle to draw current rectangle on another active image
  2. Generate Spectra from Rectangle
    1. Data Cube / Plot / Sum Spectrum within Rectangle
    2. Plots / Save As EMSA Text to save plot
  3. Acquire Rectangle Statistics
    1. Info / Rect Stats
    2. File / Save As Text to save the text popup window
  4. Acquire Single Pixel Statistics
    1. Info / Pixel Info on Click
    2. Info / Statistics on Click
    3. Info / Chart on Click
  5. Generate Threshold Masks and Plots
    1. Colorize / Threshold Slider
    2. Slide to highlight pixels in range
    3. On the threshold slider click Thresholds / Make Mask
    4. Data Cube / Plot / Sum Spect with Mask
  6. Peak Minus Background
    1. Data Cube / Slider / Peak – Background
    2. Image / Peak Minus Background