Predictive intelligence in evaluation of visual perception thresholds for visual pattern recognition and understanding

摘要

This paper will present how predictive evaluation can be applied in research on specifying perception thresholds in cognitive approach to understanding images. This approach will be based on predictive methods with cognitive inference for threshold re-production for image patterns. In particular, two basic parameters having an impact on perception thresholds evaluation, i.e. knowledge and expectations will be taken into consideration. These are the very important parameters on which perception thresholds depend. They can become lower with attempts to identify correctly or understand images by an observer, who has in-depth topical knowledge and far-going expectations concerning the image content.