基于图像分割的扭曲鼻纹识别模型仿真

Twisted Nose Pattern Recognition Based on Image Segmentation Model Simulation

鼻纹特征是当前最新研究的生物识别特征,对身份的鉴定有着重要的意义。鼻子的纹路较细小,只有在挤压过程中,才能留下纹理特征。但是,挤压过程本身对特征就形成了一定的破坏,使得特征扭曲变形,无法形成完整特征。传统的生物识别方法多是依靠完整生物特征,在鼻纹特征残缺或者挤压造成特征形变的情况下,虽然能识别出特征,但是对应过程会产生大量的干扰,造成假识别,漏识别的问题。提出采用鼻纹图像小区域分割的扭曲识别方法。针对扭曲鼻纹图像进行矩阵重组,根据重组结果进行分割,完成扭曲鼻纹图像的分类,利用自适应迭代出不同区域的识别阀值,设定识别过程中的最优阀值,最终完成扭曲鼻纹图像的识别。仿真结果表明,利用改进算法进行扭曲鼻纹图像识别,能够有效提高识别的准确率。

Nasal grain is characterized by the current biometric feature and is of great significance to the research on identity identification, on the paper proposed a nasal grain distortion recognition method based on nasal grain image small region segmentation. For twisted nose lines images, the matrix was restructured, then according to the result of restructuring segmentation, the classification of the twisted nose lines images was completed. Adaptive iteration was used to obtain the recognition thresholds of different areas, and set the optimal threshold in the process of identification, finally, the twisted nose lines image was identified. The simulation results show that the improved algorithm can effectively improve the recognition accuracy.