Based on one of the improved probabilistic neural network models - FD0 (full domain optimum) neural network, the paper presented an idea of taking the effect of the 8 neighbors of every pix element into consideration in designing the convergent domain of the network. The activation function of the network was modified, thus, endowing the network with good stability and high running speed. The recognition capabilities of FD0 network and its improved version were compared in the simulation experiment. The result proves that the improved network is especially suitable for the recognition of targets with Gaussian noise.
|Number of pages||7|
|Journal||Hongwai Yu Haomibo Xuebao/Journal of Infrared and Millimeter Waves|
|Publication status||Published - 1995 Feb|
ASJC Scopus subject areas
- Atomic and Molecular Physics, and Optics