Optical Character Recognition VMT^® OCR

Reading of characters and symbols of all kinds, such as plain text, matrix code and bar code.

The core of the system is a neuron network that can be trained to recognize characteristics with the aid of models. This network is trained to recognized model characteristics and symbols, enabling the system to read a liberal number of characteristics. 

By adding further appearance variants, the system achieves maximum recognition capability possible. Fluctuations in the surrounding conditions and varying image backgrounds may therefore be easily optimized.

The system is an equally effective tool for the end user and the OEM customer in terms of optimized production procedures, process controlling and documenting, thus reducing need for additional and follow-up work.

The unit is operated with a modern user interface that allows intuitive working. No knowledge of programming at all is required to operate the unit.

Through simple movement of the mouse, the user may call up new models and test tasks, change testing plans, or follow trained recognition.

Since operation is being kept so simple, two days of training are usually sufficient to be able to operate the system.

• Suitable also for difficult application conditions, such as changing backgrounds, etc.
• Trainable for a high variety of fonts, characteristics and symbols
• Suitable also for fast moving objects and high tact rates
• Automatic image memorizing, allowing short operation startup and time optimizing, as well as error documenting
• Accurate recognition of relevant areas through preset position definition of objects and script drafts
• The highest recognition certainty possible through use of preset knowledge of character positions or expected characters, e.g., through use of negative examples and blanking out of irrelevant areas
• The system is validable for applications in the pharmaceutical and medical industry. Conformity to 21 CFR Part 11, according to FDA Standards