Logical Designs is committed to providing the highest performance neural network solutions available to business and industry, with products that are easy to use and reliable. Specializing in neural networks since 1986, we have worked on a  wide variety of commercial and government related projects.  

Logical Designs also does consulting in areas of general software and user interface development, Windows, Java, and FORTRAN software porting to Windows. We have worked on WEB based applications including servlets, JSP and server side database programming.


Logical Designs Specialties

Abnormality Detection and Classification

Analysis of real time streams of data can be a problem when there are large numbers of parameters in the data stream.  Techniques that use bounds on parameters to trigger abnormalities usually have poor performance.  Tightening bounds to avoid missing detections usually results in too many false positive detections.

The Neural Abnormality Detection System trains neural networks on the data stream to learn what is normal.  These neural networks then becomes the detection engine for a two part system.  The second part of the system is the classification engine.  Here the detections are clustered into detection classes that can be labeled by the user as to the type of abnormality.  Previously unseen abnormalities in the data will be detected by this methodology and can trigger alarms for the user. 

Image Recognition

When an imaging application produces more images then there is time or operators to analyze them, an automated recognition system is required.  Industrial inspection and surveillance are just two of the many areas where automation is required to lower costs and raise throughput.  One example of an inspection problem is the analysis of microscope images to find and/or count item in the image.  Finding crystals or counting cells can now be performed by a computer instead of a person.

Training images are usually required for image recognition and classification problems.  These images are preprocessed by Logical Designs software to identify focus of attention points in the images. Then an operator labels these attention points into recognition classes.  Finally, customized post processing is developed by Logical Designs to score, classify or produce output counts for each image.


Logical Designs Projects 

Here is a small sample of other projects we have worked on.

Credit Scoring Project Example

Speech Recognition Project Example

Signal Processing Project Example

OCR Project Example

Financial Analysis Project Example

Robotics Project Example

Algorithm Development Example

Fuzzy Logic Project Example


Copyright 2005, Logical Designs Consulting, Inc.  All Rights Reserved