Wavelet-domain-Hidden-Markov-Models

该VC++程序是基于CUDA架构的中值滤波程序，GPU加速，并且使用了opencv函数库(The VC++ program is based on the CUDA architecture median filtering program, GPU acceleration, and use the opencv library)

两个l1准则下的噪声干扰信号压缩感知重构举例，两个例子的稀疏矩阵均为DCT矩阵，而观测矩阵分别采用单位阵和随机矩阵，有详细的步骤和使用方法，适用于初步的学习压缩感知方法。(This programme supply two examples by Compressed sensing with l1 norm. The sparse matrix of two examples are all DCT matrix and the obsever matrix are unit matrix and random matrix respectively. We give the detailed usage and steps for the programme which is useful to the beginner.)

说明：  对边缘图像作hough变换，从峰值中恢复出车道线(Right edge of the image for hough transform, recovered from a peak in the lane)

MATLAB实现图像的傅里叶变换，包括一维傅里叶变换，二维傅里叶变换(MATLAB Fourier transform of image, including the one-dimensional Fourier transform, two-dimensional Fourier transform)

蔡远利，西安交大，随机滤波与最优估计，非线性滤波，卡尔曼滤波，最小方差估计(optimal estimation)

说明：  两种基于非下采样变换图像去噪方法，效果比contourlet变换要好(Two transform based image denoising nonsubsampled method, the effect is better than contourlet transform)

可以画 花骨朵、花蕾、花朵等效果 一枝梅花等你来画哦 小游戏

Speedy GA Graphical analysis

说明：  图像检测，目标检测简单来说就是找出图像中的目标并给定目标的位置，一般用矩形框定位出目标(In image detection, target detection is simply to find out the target in the image and give the location of the target. In general, the rectangular box is used to locate the target)

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.)