Description: 采用神经网络实现手写识别的一种方法，建立Bp神经网络，采用快速训练方法，可快速完成一类相关手写字体的模式识别，识别率较高，当字体变化较大识别率降低时，可重新训练具有较强的适应性。实验证实本方法较好实现了手写字符识别，但也存在识别速度较慢，有时训练不收敛等缺点-Handwriting recognition using neural network is a way to establish Bp neural network, using fast training methods, and can be related to the completion of a class of handwriting pattern recognition, recognition rate higher recognition rate when the font changed greatly reduced, they can re- Training has strong adaptability. Experiments confirmed that this method achieved a better handwritten character recognition, but there are also identified slow, and sometimes non-convergence of the training deficiencies Platform: |
Size: 105472 |
Author:caoliang |
Hits:
Search - Handwritten Character Recognition using Neural Net