Palmprint Image Dataset with Gabor Filter Feature Enchancement

The palmprint dataset is captured on left hand. Palmprint dataset is acquired from 15 people with 5 to 8 images of each person. To increase the amount of data in each person, the raw dataset was filtered with Gabor Filter. The characteristics of the Gabor Filter are good applied to palmprint image because the image has many variations of line direction and the thickness. The palmprint dataset has 20 to 32 images each class after applying the Gabor Filter. The author trains the palmprint dataset using the Convolutional Neural Network method.

该掌纹数据集系从左侧手掌采集而来。本数据集由15位个体提供，每人提供5至8张图像。为增加每位个体的数据量，原始数据集经Gabor滤波器过滤处理。鉴于掌纹图像具有丰富的线方向和厚度变化，Gabor滤波器的特性在此图像处理中尤为适用。经过Gabor滤波器处理后，每个类别的掌纹数据集包含20至32张图像。作者采用卷积神经网络方法对掌纹数据集进行了训练。