Image processing and recognition technology

If you are curious to know the underlying technologies, you can get the information right here.

Picollator image search and indexing system is a true hi-tech solution. It is based on two main principles: intelligence and architecture.

The intelligence is needed for image processing, object detection and face recognition. We made significant efforts to simulate mathematically how people represent the environment in their brains. We use non-linear filters and adaptive learning technique, which allows Picollator to distinguish different objects.

The system recognizes the detected object. If it is a face, Picollator has to decide whose face it is. It knows some faces from its previous work and attempts to select the most similar images to the new one. This process is mostly based on artificial neural network models together with novel information representation concept developed by Recogmission.

Finally the system indexes the images and makes the non-textual dictionary. So, the system performs just like people do in their brains.

The architecture is smart as well. Picollator software consists of several components, which can be installed independently on different computers across the network or on a single PC. The components work in the manner of GRID computing paradigm and can process information using multiple resources according to the tasks. Even if the component cannot complete the task, other components will attempt to solve the task or wait until another running component takes the task by itself.

The system is absolutely stable and adaptive to changing computing resources and network configurations. The architecture is realized as the computing platform and may be implemented as a separate product.

Our technology is implemented both in Picollator Online and in Picollator image search and indexing system.

If you have questions about our image search and indexing software, contact us.