GENIAL

Section: User Commands (1)
Updated: 27 August 1991
Index Return to Main Contents

NAME

genial - (GENeral Image AnaLysis): tool for quantitative analysis of HIPS images

SYNOPSIS

genial

DESCRIPTION

genial is an interactive image analysis environment for the X Window System. It provides a number of tools for exploration, quantitative analysis, and annotation of images of arbitrary size and resolution. In addition, genial maintains a log of operations in the image header so that all analysis and annotation can be kept with the original image.

OPTIONS

Standard XView options

CONTROL WINDOW LAYOUT

The window titled ``Genial Control'' is displayed immediately upon startup. It is divided into two parts: The upper half is used to control genial's operation, and the bottom part displays status information. The status information consists of four lines of instructions, and a bottom line which shows the image window status. The (x,y) coordinates of the cursor position within the image are displayed, along with the pixel value under the cursor. The right hand portion of this line is used for gross features of the image. These features are the geometry, dynamic range, and pixel format (in that order).

The following describe the different parts of the top half of the control window:

``quit'' button -: allows you to quit the program.
``Image'' items -: allow you to load images, change image display parameters, and traverse the frames of a multiple frame image sequence.
``Function'' items -: allow you changes the analytical or annotation function. Also allow you to evaluate or clear the current function and bring up additional function options.
``Log'' items -: allow you to control whether or not automatic logging of functions takes place.
``Regions'' items -: allow you to change the region types and the line drawing mode.
``done'' button -: button used to signal completion of an operation in certain parts of the program.

USAGE

After the control window appears, a file is loaded by using the ``file'' button. This button displays a pop up window. Type the name of a HIPS image file into the ``Filename:'' text field and press return. This will load the image into memory and display it on the screen.

After the image is loaded, the user is free to use any of the analysis functions. To do this, select an analysis function from the ``Function:'' selector. Then select a region type from the ``Type:'' selector. Then choose the control points in the image which define the region. Control points are added by pressing the left mouse button. A cross will be displayed in the standout color over the image to give a visual indication that the point has been selected. Points may be interactively ``moved'' using the middle mouse button, and may be deleted using the right button.

After you have selected a region to your liking, hit the ``eval'' button in the control window. The analytical function will be evaluated based on the region you specified, and the results of the analysis displayed. In addition, the function will be entered into the log, and a log ID number will appear next to the first point you selected.

You may repeat this process indefinitely. You may select a new analytical function or use the same function. There is no limit to the number of invocations of the analysis functions.

At any point, you may save your work by using the file button and typing the name of a new file in the save filename. This will save both the image and a log of all operations. If this file is later loaded, all functions in the log will also be displayed automatically.

IMAGE ITEMS

``file'' button -: pops up a window for file I/O. A HIPS image is loaded by typing the name of the file in the ``Filename:'' text field under the word ``Load:''. The image (with the log) can be saved by typing the file name into the ``Filename:'' text field under the word ``Save:''.
``display'' button -: pops up a window for modification of the display image. All analysis will still happen on the original image dataset. There is a slider for contrast enhancement, and two numeric text fields for changing the colormap scaling. In addition, there is a ``shrink factor'' selection item. The image width and height will be divided by this shrink factor. genial will keep track of the mapping between the display image and original image so that coordinate conversion happens transparently. The smaller image is computed by local averaging, in order to arrive at a more suitable approximation of the original image.
``<'' and ``>'' buttons -: are used to traverse through the frames of a multiple frame sequence of images.

FUNCTION ITEMS

``Function:'' selector -: right button pulls down a stack to select the type of analytical function.
``eval'' button -: evaluates a function after a region has been selected. Note that a region must be selected before a function can be evaluated.
``clear'' button -: clears away the current function. The current function can be easily identified, as it has the highest numbered log ID. During the middle of selecting points for the current region, the clear button will clear away those points.
``options'' button -: brings up an options window for the type of analytical function displayed by the ``Function:'' selector.

LOG ITEMS

``Auto Log'' checkbox -: toggles whether analysis functions will automatically be stored in the log. The default is on, and it is probably a good idea to leave auto logging on.

REGIONS ITEMS

``Type:'' selector -: right button brings up a stack of valid region types. The set of valid region types is based on the current function type. Obviously certain functions (e.g. zoom) can only operate on certain types of regions (in this case, rectangles).
``Line Drawing Mode:'' selector -: right button brings up a stack that allows the user to change the line drawing mode. ``Click and Click'' mode allows the user to just click on the two endpoints for the line. ``Click and Drag'' mode requires the user to click on the start point and "drag" the line to the endpoint. ``Least Squares Fit'' mode allows the user to click on an arbitrarily large number of points. genial will select the best line through the points. Note: ``Click and Drag'' mode is abysmally slow in this implementation.

ANALYTICAL FUNCTIONS

trace

Traces are simple graphs of the gray values along linear regions (lines or splines) of an image. A simple control panel is also provided to allow for options such as an averaging radius which smooth the graph by displaying the average gray value at each point along the region.

In addition, there are two other region types for traces: Boxes, which give the average pixel value along a line parallel to the width of the box. Another region type, "Parallel Path" gives the average value pixel value on a line parallel to several reference lines which the user supplies.

Traces may also be "integrated". That is, it is possible to do a separate graph of the area under specific portion of the trace graph. To select regions for integration, click the left button to select the start and the right button to select the end markers in the trace display window. When all regions are selected, click on the ``done'' button in the main control window.

angle measure

Allows measuring angles between two lines.

distance

Function which measures the linear distance, in pixels, between two points.

zoom

Zoom allows the user to interactively zoom various rectangular regions of an image. The zoomed region is displayed in a separate window, and the user may increase and decrease the zoom factor interactively. In addition, the user may interactively move the region which is zoomed. This is done through an experimental interface, log_chreg(), which changes the region associated with a particular log entry. This is discussed further in section 7.

display control

A display control window is provided which allows the user to change the scaling and contrast of the colormap and select a shrink factor to facilitate analysis of huge images. These are not actually implemented as function modules because they modify the colormap and the original image which are global resources, and display enhancement is not really a "quantitative analysis" function.

histograms

Histograms are very useful for finding how and where the real information is in an image. Histograms operate on closed regions of an image (polygons, closed splines and rectangles). The histogram window displays the intensity values on the horizontal, and a pixel count on the vertical. The height of any single bar in the graph represents the number of pixels in the selected region with that particular intensity value. This information can be particularly useful for scaling the colormap. The colormap is, by default scaled evenly between the minimum value and the maximum value in the image. However, the actual distribution of useful data in the image might be in a much smaller region, and the extrema of the data values might just be artifacts. A histogram can graphically illustrate this. For example, if 90% of the data in a 16-bit image is located in the gray value range 200-250, but there were a very pixels with the value 3 and a very pixels with the value 32,564, the data values between 200 and 250 would all be mapped to the same colormap entry. Based on seeing this huge disparity in the histogram, the user could rescale the colormap so that the data values between 200 and 250 cover achromatic gray from black to white.
In addition, the histogram options window displays a palette of 6 colors. The user may select a subrange of the histogram window, select a color from the palette, and all pixels within the selected range in the original image will be illuminated with the selected standout color. This, also, is useful in analyzing just where and how the information in an image is distributed.

BUGS

The ``done'' button is very messy from a UI point of view. It would be nice if there were some better way.

The log should be stored in the extended parameters section of the HIPS 2 header, not in the sequence description field.

There are probably plenty of others.

AUTHOR

Antony A. Courtney,
Imaging Technologies Group, Lawrence Berkeley Laboratory and
Computer Science Department, Trinity College, University of Dublin, Ireland.
Email: antony@george.lbl.gov

This document was created by man2html, using the manual pages.
Time: 01:44:48 GMT, December 05, 2024