【摘要】 文章将智能控制与模式识别方法相结合,提出了一种仿人识字行为的图像汉字智能识别机;采用决策控制级、模式识别级和反馈校正并行式结构。决策控制级选择最佳的识别策略和参数,模式识别级进行多分类器结果融合识别,并根据对模式识别级输出的可信度判断结果进行反馈,校正识别方法和参数,最终提高汉字的识别效率。更多还原

【Abstract】 An intelligent recognizing machine of Chinese character images is presented in this paper.It combines intelligent control with pattern recognition.A parallel structure consisting of the decision-making level,the pattern recognition level and the feedback level is adopted.At the decision-making level,the best recognizing strategies and parameters are chosen.At the pattern recognition level,multi-classifier results fusion is carried out.At the feedback level,different results of recognition are fed back so as to revise the recognizing methods and parameters.This structure can increase the recognition efficiency.更多还原