A hierarchical interaction architecture for pattern recognition
Myung Won Kim, Gowang Lo Lee, Jaehoon Kim, Chaedeok Lim · 2002
Summary form only given. Proposes a hierarchical interaction neural network (HINT) model for complex pattern recognition. HINT recognizes handwritten Hangul characters in a context-dependent way. HINT consists of two subnets, FD-net and HC-net, which detect Hangul alphabets and implement Hangul character composition rules, respectively. The underlying theme is that complex pattern recognition involves a highly interactive process in a hierarchy of different functional processors. The proposed model suggests that hierarchical interaction is an efficient architecture for complex pattern recognition. Since the model supports high modularity, it is easy to implement a HINT-like neural network for a given problem.>