基于深度卷积网络的主掌线纹区域的提取

Extracting Main Lines Area on Palm Based on Deep Convolutional Network

鉴于生物身份识别的普及,掌纹识别系统越来越要求对环境和图像质量的鲁棒性。为了解决这个问题,提出了一个新奇的基于深度学习的方法。通过逐渐的由粗到细地定位主掌纹线所在的区域,能够较好地适应大部分的实际数据和环境。最后,通过在测试集上的测试,该算法取得了较为良好的效果。

As the biometric identity have been used widely,more and more palm-print recognition systems require Robust localization within various environment and image quality.By using a novel method which gradually localize main line on palm from coarse to fine,this could help to adapt most case in the actual conditions and palm-print database or system.