基于ELM的多结构变形QR码分类校正研究

针对不同曲面上QR码变形多样化、识别率低的问题，考虑到极限学习机(ELM)对大量数据的快速分类能力，提出了一种基于ELM的多结构变形QR码分类校正算法。在欧式距离量化变形特征后，运用ELM算法把变形结构分为平面变形、半曲面变形和全曲面变形三类，并利用不同分类系数改进QR码坐标透视变换算法，得到校正坐标值。实验结果表明，此方法不仅提高了QR码在曲面上的校正准确率，而且通过分类提高了曲面变形和平面变形QR码的校正速度。

Multi-structural QR Code Recognition Based On ELM

Since the traditional QR code recognition methods fail to fulfill the recognition task on cylindrical and spherical deformed QR code cases, a novel QR code recognition algorithm based on extreme learning machine is presented in this paper. In this method, Euclidean distance of feature points are calculated as deformation characteristics. Extreme learning machine is employed to classify the deformation type. It determines the transform coefficients of each type, which are used for coordinates mapping. The real coordinates are obtained by perspective mapping with different transform coefficients, thus the code is reconstructed. Experimental results demonstrate that our proposed method hold a satisfactory performance on both of recognition rate and time consuming.