2006011437_HW_2

自编语谱图函数，可实现基本的spectrogram功能(Self compiled spectral function, which can implement basic spectrogram functions)

基于TensorFlow图像处理的小测试，里面的代码自己尝试过都可以运行。(Based on the TensorFlow image processing test, the code can run if it has tried.)

使用GDI 进行动态会制多叉树代码，例子是使用FP-Tree算法生成的一个多叉树进行动太绘制

用哈夫曼算法对图像进行压缩， 1.颜色模式转换及采样； 2．DCT变换（离散余弦变换）； 3．量化；4．编码（有算术编码和哈夫曼编码两种，这里采用哈夫曼编码）

关于分水岭图像分割的源码，需要的可以下来用(watershed image segmentation on the source need to be down with the)

Jianchao Yang 的基于稀疏表示的单幅图像重建的原始代码，先将高低训练图像分块，再将块训练成高低字典，将测试图像映射到低字典上，得到系数，再乘以高子典就得到最后的图像。对学习超分辨率同学的参考作用很大。(This is the original matlab code for super resolution by Jianchao Yang 。The method is sparse represent based on the overcomplete dictionary。)

自己编写的图像处理函数，包括绘制直方图，直方图规定化，直方图均衡化，具体算法参阅章章毓晋. 图像工程（上册）——图像处理（第二版）[M]. 清华大学出版社(Write your own image processing functions, including the plot histogram, histogram specification, histogram equalization, the specific algorithm see chapter Yu-Jin. Image Engineering (on)- Image Processing (Second Edition) [M]. Qinghua University Publishing society)

一个改进的卡尔曼滤波器 用来进行运动目标的跟踪 (An improved Kalman filter used for tracking moving targets)

关于图像分割中分水岭的一个代码，能够实现相应功能，拿出来与大家共享(Image segmentation in the watershed on a code, can achieve the corresponding functions, and show to share with you)

对图像进行二维离散小波变换, 变换级数大于等于3级,然后进行阈值化处理(阈值约为10左右),再统计系数中0的个数（百分比表示）并进行重构, 最后计算重构图像的峰值信噪比(PSNR). (Two-dimensional discrete wavelet transform image transformed series is greater than or equal to 3, and then processing thresholding (threshold of about 10 or so), then statistical coefficient in the number (expressed as a percentage) and to reconstruct the final calculation reconstructed image of the peak signal-to-noise ratio (PSNR).)