基于HMM与BP神经网络的物体识别算法

Object recognition algorithm based on HMM and BP neural network

通过对现有物体识别算法的研究,提出了一种将HMM模型与BP神经网络相结合的方法进行物体识别,将HMM模型的最佳输出状态作为BP神经网络的输入。所设计的模型包括HMM模型设计、BP神经网络结构设计、图像库中图像设计。实验证明该算法对物体的识别率较高,鲁棒性强,训练好的模型具有比较强的识别能力,具有一定的实用价值和广阔的发展前景。

Based on the existing object recognition algorithm is proposed,and the model will be HMM and BP neural network method object recognition,the best output state HMM model as the BP neural network of input.The model includes HMM designed and BP neural network model design,structural design,image database design.Experiment results show that the algorithm is high rate of objects,robust,trained model has stronger ability to identify,has certain practical value and broad development prospects.