CAJ | 학술논문

为了有效地对印章图像进行分类，提出了一种基于 Krawtchouk 矩和 RBF 神经网络的印章图像分类识别方法。首先提取标准印章图像的 Krawtchouk 矩不变量，将其作为 RBF 神经网络的原始训练样本；然后提取全部待鉴印章图像的 Krawtchouk 矩不变量，将其作为 RBF 神经网络的输入量进行分类识别。实验结果表明，使用 Krawtchouk 矩来描述印章图像的特征并通过 RBF 神经网络来对其进行分类识别的方法十分有效，与同类的 Brushlet-RBF 法和 KPCA-RBF 法相比，其识别率更高，且更准确。
위료유효지대인장도상진행분류，제출료일충기우 Krawtchouk 구화 RBF 신경망락적인장도상분류식별방법。수선제취표준인장도상적 Krawtchouk 구불변량，장기작위 RBF 신경망락적원시훈련양본；연후제취전부대감인장도상적 Krawtchouk 구불변량，장기작위 RBF 신경망락적수입량진행분류식별。실험결과표명，사용 Krawtchouk 구래묘술인장도상적특정병통과 RBF 신경망락래대기진행분류식별적방법십분유효，여동류적 Brushlet-RBF 법화 KPCA-RBF 법상비，기식별솔경고，차경준학。
In order to effectively categorize the seal image, a method of classification and recognition of seal image based on krawtchouk moment and RBF neural network is proposed. Firstly, krawtchouk moment invariants for part of the standard seal images are calculated to train the RBF neural network. Then characteristics of all the questioned seal images are applied to recognize. Results show that Krawtchouk moments as a feature description of the seal image by RBF neural network to classify the identification method is very effective, its recognition rate is higher, and more accurate, compared to the same kind of Brushlet-RBF and KPCA-RBF.