MODULAR EXPERT NETWORK APPROACH TO HISTOGRAM THRESHOLDING

The problem of histogram thresholding is tackled using a modular exper t network. The modular expert network is a network of expert modules m odulated by a gating network. The expert modules incorporate individua l experts' opinions on the thresholding problem. The difficult task of integration of conflicting experts' opinions is achieved through a tr aining of the gating network using backpropagation. The resulting netw ork achieves accurate modeling of the solution mapping through the eff icient combination of existing experts. Experimental results show the superior performance of the modular network over classical algorithms. In particular, a near-optimal solution was shown to be achievable usi ng a small training set Application to a real-world biomedical cell se gmentation problem is also given, (C) 1997 SPIE and IS&T.