home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Amiga Mag HDD Backup
/
Amiga Mag HDD Backup.zip
/
Amiga Mag HDD Backup
/
Alexander.img.bin
/
Alexander.img
/
*********10.11
/
AC 10.11 Archive.sit
/
ChestNetPictures.ascii
< prev
next >
Wrap
Text File
|
1995-11-16
|
2KB
|
74 lines
Explanation of Chest Net Images
FIGURE 1
The NeuroPro (MegageM) icon helps explain the neural network
concept. The neural network is given data in the Input Layer and then
told the the meaning of that data by information placed in the Target
Layer. Training the system requires that Input and Target pairs be
presented over and over again. The Hidden Layer, which has
connections to every cell in both layers, adjusts the strength of its
connections until the Output Layer matches the Target Layer for each
Input-Target pair. Once the network, i.e. Hidden Layer, has been
trained, new data can be placed in the Input Layer. The network will
then respond with an answer in the Output Layer.
FIGURE 2
To see which disease your x-ray corresponds to, just check the
boxes that describe what you see. Using ARexx, Chest Net will send
the data to NeuroPro and get its opinion.
FIGURE 3
Before you can get a consultation from your Amiga, you will first
need to create a file that contains descriptions of typical chest
diseases. After you check the appropriate boxes for each disease,
Chest Net will do this for you.
FIGURE 4
Chest Net can tell you what diseases it knows about, allow you to
add to its list of diseases, train the neural network and even tell
you which disease your patient most likely has -- provided you can
describe the x-ray findings.
FIGURE 5
Chest Net can access images you have stored on disk so you can
compare them with those of your patient. This CT scan shows what you
might expect to find in Alveolar Proteinosis.