HOME - PATTERN EXPERT HOME - PATTERN EXPERT
     
  
   deutsch  

Screenshots


Organizing the Knowledge Base

- Organizing the Knowledge Base



Knowledge Base Organizer : By Hand
Administration of Knowledge Bases.
Exchange your Knowledge Bases
via export and import.
Organize spectra and classes. Supplement the spectra with additional user defined information.
Organizer : By Directories Organizer : By Table
Import spectrum files
simply by Drag&Drop.
Use a tabular catalog of your spectra for import and assignment of classes.



Visualization

- Visualization



Visualization : Overlay Visualization : Stacked Classes
Multiple spectrum view
(here: all in one diagram).
Class means and standard deviations
(here: in separate diagrams).
Visualization : 2.5D View Visualization : Gel View
2.5D View - Perspective view of the spectra in class common colors. Top view of spectra ("GelView") -
classes indicated by color codes.



Configuration

- Configuration



Configuration : Baseline
Configure the preprocessing steps -
These are applied to all spectra automatically.



Training Process

- Training Process



Learning : Information Learning : Assignment Table
Information on classes and spectra as well as the classification system's state, presented in a clear manner. Detect spectra with a "suspect"
training process behavior (outliers),
and examine them thoroughly.
Learning : Class Table Learning : ROI Viewer
The class assignment table reveals classes, which are frequently confused. Representation of those spectral regions, which are used to
distinguish the classes.



Recognition

- Recognition



Recognition : Unused Reference Spectra Recognition : Analysis and Retrieval
Classification of new unknown spectra, or of spectra previously excluded from training by the user. Verify the results and search for the most similar reference spectra providing evidence for class assignments.



Simulation of "Real-Life" Conditions

- Simulation of "Real-Life" Conditions



External Cross Validation : Block Overview External Cross Validation : Single Block Results
The fully automatic "External Cross Validation" performs a simulation of the conditions which arise in practical applications. Single block analysis. Analyze the spectral features found to be useful and the calculated confidence values for correct and false class decisions.

External Cross Validation : Feature Comparison
Comparison of relevant features as determined for the individual blocks.



 
 

        : Home :: Company :: Competencies :: Imprint :: Sitemap :