摘要
目前关于刺绣的研究基本上都是偏于从花版文件生成针位文件的算法,而识别方面的算法尚为空白。在已有的商业软件中相关功能仍需较多的人工干预,并且识别率不高。分析不同刺绣格式的走针特征,基于神经网络设计分块算法和轮廓识别算法。实验表明该算法可行,速度快,可以为刺绣软件相关模块的设计提供参考。
Embroidery are basically biased from the flower version file generated pin file algorithm, the identification algorithm is still blank. Related functions in the existing commercial software still need to be more manual intervention, and the recognition rate is not high. Analyzes the characteristics of the different formats of embroidery go pin, a block algorithm and contour recognition algorithm based on neural network design. The experiment shows that the algorithm is feasible, and speed. And it provides a reference for the design of the embroidery software module.
基金
广东省大学生创新基金
关键词
刺绣针位
神经网络
数学模型
模式识别
Embroidery Needle Position
Neural Network
Mathematical Model
Pattern Recognition
