Patterns in Our Daily Life

The behavior of each organism in nature forms a profile which makes it distinguishable from others. Distinct attributes can also be observed in physical phenomena’s surrounding our daily life. We call distinctive profiles around us patterns. Pattern recognition deals with the automatic identification of regularities using computer technologies. In this article, we will focus on how pattern recognition and automation makes our lives easier.

Pattern recognition has three main categories. The first and main application area is called classification. Character recognition is a typical example of classification. Imagine using your PDA to write down a unique idea during your lunch time, or save the phone number of a person you met. Your PDA basically recognizes the handwriting and classifies your strokes as a letter in alphabet during this process. It performs a very cumbersome task considering that each person has a unique writing style. If you are not impressed by character recognition, we have gene structure analysis as a classification application of pattern recognition. Current developments in genetic research enable us estimate the possibility and risk of having a certain disease. Just Google 23andMe, and see how a personal genetic analyzer became the invention of the year. This pattern recognizer associates your genes with diseases and reveals the correlation scores. These correlation scores are obtained after training large data sets across the nation. Remember the sci-fi movie, Gattaca, where an ambitious individual gets rejection for a job due to imperfections in his gene structure. It doesn’t sound like sci-fi any more right?

The second major application of pattern recognition is novelty detection, where an unexpected behavior is automatically detected. Google’s new release of “Flu Trends” enables catching an epidemic at early phase using the words that are entered on search engine. The third main application of pattern analysis is regression, which is used for estimating a function from noisy data. Regression is adopted by the financial industry to estimate the stock market’s behavior and other financial trends.

The common attribute of all pattern recognition tasks is to predict possible outcomes of events, which leads to automation. Automation allows extracting useful information out of complicated inputs, i.e. gene structures or distorted handwritten letters. At LVS we realize the importance of automation in print inspection and bar code verification. We built automatic 1-D & 2-D bar code detection and automatic character extraction with this spirit. Our automated system extracts bar codes and associated alpha-numeric sequences without any interference. We believe that print waste will be reduced to great extent as automation becomes vital in print inspection.