ClockDemo

说明：  利用全卷积神经网络，实现图像的语义分割，基于tensorflow的keras可以直接运行(Using the full convolutional neural network to achieve semantic segmentation of images, keras based on tensorflow can be run directly)

RINGSDK包含界面库和图象库,我编译过，可以使用(RINGSDK is a interface lib )

本件包含四种二维高通滤波器，他们分别是基于汉宁窗，布莱克曼，凯泽窗和巴特利特窗设计的二维高通滤波器。(This piece contains four two-dimensional high-pass filter, they are based on Hanning, Blackman, Kaiser window design windows and Bartlett two-dimensional high-pass filter.)

图像检测的算法，Faster R-CNN算法，先对整张图像进行卷积计算，然后通过感兴趣区域池化层(RoI Pooling Layer)将选择性搜索算法推荐出来的候选区域和卷积网络计算出的特征映射图进行融合，得到候选区域对应的特征矢量，这种共享卷积计算的操作极大地减少了卷积计算的次数。而且这些特征矢量的维度统一，方便后续的分类工作。通过感兴趣区域池化层处理卷积特征，并将得到的特征送往两个并行计算任务进行训练，分类和定位回归。通过这些方法和改进的框架，Fast R-CNN 用更短的训练和测试时长，取得了比 R-CNN 更好的效果(Faster R-CNN algorithm first convolutes the whole image, then fuses the candidate regions recommended by the selective search algorithm and the feature mapping maps calculated by the convolution network through the RoI Pooling Layer to get the corresponding feature vectors of the candidate regions, which greatly reduces the number of convolution calculations. Moreover, the dimension of these feature vectors is unified, which facilitates the subsequent classification work. The convolution feature is processed by the pooling layer of the region of interest, and the obtained feature is sent to two parallel computing tasks for training, classification and positioning regression. Through these methods and improved framework, Fast R-CNN uses shorter training and testing time and achieves better results than R-CNN.)

说明：  首先对指纹图像进行预处理，包括图像的分割，增强，二值化，细化。对细化后的图像进行特征点的提取，最后通过细节特征匹配完成指纹的识别。(Firstly, the fingerprint image is preprocessed, including image segmentation, enhancement, binarization and thinning. The refined image is extracted by feature points, and the fingerprint recognition is completed by minutiae feature matching.)

数学建模，动态模糊与模糊图像处理，维纳滤波方法。(Mathematical modeling, dynamic fuzzy and fuzzy image processing, Wiener filtering method.)

基于预测误差周期性的图像重采样检测程序。(Image resampling detection based on the periodicity of prediction error)

一个盲去卷积程序，可以实现去除图像的各类模糊，同时可进行相关参数的自适应调整。(A blind deconvolution process can be achieved to remove all kinds of image blur, while the relevant parameters can be adaptive.)

DCT域图像水印算法，并根据携密载体提取秘密信息(DCT domain image watermarking algorithm, and extract secret information according to the carrier.)

主要有图像的聚类分割和Ncut算法的彩色图像分割，对于新人有着很好的效果(There are clustering image segmentation algorithm and Ncut color image segmentation, it has a good effect for newcomers)