The Boolean Neural Network (BNN) has been successfully applied in many image processing tasks. In this paper, a method to incorporate rules into the BNN is proposed. The concept of ROA (Radius of Attraction) is used to incorporate more complex rules into the BNN which are difficult and inefficient to incorporate in traditional production systems. Then, a Rule-Based BNN approach is proposed as an autonomous approach for image understanding. This approach integrates unsupervised image segmentation, feature extraction, image labeling, and a task-specific knowledge-base to extract information from images. The approach was applied to the analysis of colored driver licenses (Kabuka, Waly, and Sauer, 1995). The target was to extract the driver personal information. Results of testing are provided in the paper.
|Title of host publication||Intelligent Engineering Systems Through Artificial Neural Networks|
|Number of pages||5|
|State||Published - Dec 1 1996|
ASJC Scopus subject areas