cv相关博客文章

　　收藏些图像处理，机器学习，深度学习方面比较不错的文章，时常学习，复习和膜拜吧。。。

图像方面（传统CV）：

1. SIFT特征

https://www.cnblogs.com/wangguchangqing/p/4853263.html

http://shartoo.github.io/SIFT-feature/?FbmNv=5d9f3d0c8ca5090a

https://blog.csdn.net/u010440456/article/details/81483145

2. HOG特征

http://shartoo.github.io/HOG-feature/?FbmNv=5d9f3d48e0647071

https://senitco.github.io/2017/06/10/image-feature-hog/

https://www.cnblogs.com/aoru45/p/9748481.html

https://zhuanlan.zhihu.com/p/40960756

3. 图像金字塔

http://shartoo.github.io/image-pramid/?FbmNv=5d9f3d6e990e41bb

https://zhuanlan.zhihu.com/p/80362140?utm_source=wechat_session&utm_medium=social&utm_oi=71873182302208

https://zhuanlan.zhihu.com/p/32815143

https://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_imgproc/py_pyramids/py_pyramids.html

4. Haar特征

http://shartoo.github.io/img-haar-feature/

https://senitco.github.io/2017/06/15/image-feature-haar/

https://juejin.im/post/5b0e6f83f265da0910791a38

https://blog.csdn.net/zouxy09/article/details/7929570

 5.Harris角点

 https://www.cnblogs.com/ronny/p/4009425.html

https://senitco.github.io/2017/06/18/image-feature-harris/

https://zhuanlan.zhihu.com/p/42490675

https://zhuanlan.zhihu.com/p/36382429

机器学习方面：

1. Linear Regression

https://yoyoyohamapi.gitbooks.io/mit-ml/content/%E7%BA%BF%E6%80%A7%E5%9B%9E%E5%BD%92/articles/%E7%BA%BF%E6%80%A7%E5%9B%9E%E5%BD%92%E4%B8%8E%E6%A2%AF%E5%BA%A6%E4%B8%8B%E9%99%8D.html

https://zhuanlan.zhihu.com/p/45023349

2. Logistic Regression

https://chenrudan.github.io/blog/2016/01/09/logisticregression.html

https://www.jiqizhixin.com/articles/2018-05-13-3

https://zhuanlan.zhihu.com/p/28408516

3.Neutral Network

https://clyyuanzi.gitbooks.io/julymlnotes/content/dl_nn.html

https://www.cnblogs.com/subconscious/p/5058741.html

神经网络损失函数（loss function）:

4. 回归和正则化(Regression and Regularization)

https://www.zhihu.com/question/20924039

http://studyai.site/2016/09/04/%E6%96%AF%E5%9D%A6%E7%A6%8F%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E8%AF%BE%E7%A8%8B%20%E7%AC%AC%E4%B8%89%E5%91%A8%20(4)%E6%AD%A3%E5%88%99%E5%8C%96%EF%BC%9A%E8%A7%A3%E5%86%B3%E8%BF%87%E6%8B%9F%E5%90%88%E9%97%AE%E9%A2%98/

https://zhuanlan.zhihu.com/p/29957294

线性回归，逻辑回归和神经网络带正则化的损失函数：

 

 正则化项能减缓梯度的变化：

 5. SVM(support vector machine)

拉格朗日乘子法

https://www.zhihu.com/question/38586401/answer/457058079?utm_source=wechat_session&utm_medium=social&utm_oi=71873182302208&hb_wx_block=0

对偶问题：

https://zhuanlan.zhihu.com/p/31131842?utm_source=wechat_session&utm_medium=social&utm_oi=71873182302208

KKT条件：

https://zhuanlan.zhihu.com/p/26514613?utm_source=wechat_session&utm_medium=social&utm_oi=71873182302208

SVM原理：

https://webcache.googleusercontent.com/search?q=cache:0bibeWe0EQYJ:https://raw.githubusercontent.com/liuzheng712/Intro2SVM/master/Intro2SVM.pdf+&cd=13&hl=zh-CN&ct=clnk&gl=us

https://zhuanlan.zhihu.com/p/24638007?utm_source=wechat_session&utm_medium=social&utm_oi=71873182302208

https://www.jiqizhixin.com/articles/2018-10-17-20

https://www.cnblogs.com/leftnoteasy/archive/2011/05/02/basic-of-svm.html

https://wizardforcel.gitbooks.io/dm-algo-top10/content/svm-1.html

https://blog.csdn.net/v_JULY_v/article/details/7624837

支持向量机的表达式，拉格朗日函数，对偶问题和KKT条件：

软间隔支持向量机的表达式，拉格朗日函数，对偶问题和KKT条件：

支持向量机非线性化的核函数：

SVM使用代码（sklearn包）：（线性svm，和采用核函数的非线性SVM）

https://zhuanlan.zhihu.com/p/37640777?utm_source=wechat_session&utm_medium=social&utm_oi=71873182302208

 SVM的python实现： https://blog.csdn.net/laobai1015/article/details/82763033

6. kmeans算法

https://yoyoyohamapi.gitbooks.io/mit-ml/content/KMeans/articles/K-Means%E7%AE%97%E6%B3%95%E6%AD%A5%E9%AA%A4.html

https://www.csuldw.com/2015/06/03/2015-06-03-ml-algorithm-K-means/

https://bainingchao.github.io/2018/09/19/%E4%B8%80%E6%AD%A5%E6%AD%A5%E6%95%99%E4%BD%A0%E8%BD%BB%E6%9D%BE%E5%AD%A6K-means%E8%81%9A%E7%B1%BB%E7%AE%97%E6%B3%95/

https://www.cnblogs.com/pinard/p/6164214.html

 k-Means++

https://zhuanlan.zhihu.com/p/32375430

kmeans和kmeans++ python代码实现：

https://github.com/silence-cho/cv-learning/blob/master/week4/assignment.py 

https://github.com/ViperBling/CV_Course/blob/master/Week5/K-Means%2B%2B/K-Means.py

7.KNN（k近邻）算法

https://coolshell.cn/articles/8052.html

https://www.cnblogs.com/ybjourney/p/4702562.html

8.决策树 (Decision tree)

https://bainingchao.github.io/2018/09/19/%E4%B8%80%E6%AD%A5%E6%AD%A5%E6%95%99%E4%BD%A0%E8%BD%BB%E6%9D%BE%E5%AD%A6%E5%86%B3%E7%AD%96%E6%A0%91%E7%AE%97%E6%B3%95/

https://www.csuldw.com/2015/05/08/2015-05-08-decision%20tree/

https://lotabout.me/2018/decision-tree/

https://blog.csdn.net/xbinworld/article/details/44660339

信息增益：

信息增益率：

基尼指数：

 ID3（信息增益）和C4.5（信息增益率）：https://zhuanlan.zhihu.com/p/26760551?utm_source=wechat_session&utm_medium=social&utm_oi=71873182302208

基尼不纯度(基尼指数)：https://www.zhihu.com/question/296781126/answer/508112100?utm_source=wechat_session&utm_medium=social&utm_oi=71873182302208&hb_wx_block=0

sklearn实现决策树：https://www.v2ex.com/amp/t/544062

 9.其他算法

AdaBoost：

　　https://www.cnblogs.com/pinard/p/6133937.html

　　https://blog.csdn.net/guyuealian/article/details/70995333

LDA(隐式狄利克雷分布):  https://github.com/endymecy/spark-ml-source-analysis/blob/master/%E8%81%9A%E7%B1%BB/LDA/lda.md

朴素贝叶斯：https://www.cnblogs.com/leoo2sk/archive/2010/09/17/naive-bayesian-classifier.html

                    https://zhuanlan.zhihu.com/p/26262151

　　　　　  python实现：https://zhuanlan.zhihu.com/p/32183117?utm_source=wechat_session&utm_medium=social&utm_oi=71873182302208

深度学习方面

1. overfit/underfit (过拟合和欠拟合)

https://marian5211.github.io/2018/03/08/%E3%80%90%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E3%80%91%E8%BF%87%E6%8B%9F%E5%90%88%E3%80%81%E6%AC%A0%E6%8B%9F%E5%90%88%E5%8F%8A%E5%85%B6%E8%A7%A3%E5%86%B3%E5%8A%9E%E6%B3%95/

https://zh.d2l.ai/chapter_deep-learning-basics/underfit-overfit.html

https://zhuanlan.zhihu.com/p/29707029

 2. bias and variance (高偏差和高方差)

https://www.jianshu.com/p/a585d5506b1e

https://www.cnblogs.com/hutao722/p/9921788.html

http://nanshu.wang/post/2015-05-17/

http://www.voidcn.com/article/p-tqoebcaa-dq.html

3.卷积

 

 

 反卷积(Deconv / Transposed Convolution / Fractionally strided conv):

https://www.zhihu.com/question/48279880?sort=created

https://www.zhihu.com/question/48279880/answer/838063090

4. Gradient vanishing and explosion （梯度消失和梯度爆炸）

https://blog.csdn.net/qq_25737169/article/details/78847691

https://jimmy-walker.gitbooks.io/tensorflow/%E6%A2%AF%E5%BA%A6%E6%B6%88%E5%A4%B1%E5%92%8C%E6%A2%AF%E5%BA%A6%E7%88%86%E7%82%B8.html

https://hunto.github.io/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/2018/07/17/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E4%B8%AD%E6%A2%AF%E5%BA%A6%E6%B6%88%E5%A4%B1%E4%B8%8E%E6%A2%AF%E5%BA%A6%E7%88%86%E7%82%B8%E9%97%AE%E9%A2%98%E8%AF%A6%E8%A7%A3.html

https://codertw.com/%E7%A8%8B%E5%BC%8F%E8%AA%9E%E8%A8%80/583004/

https://zhuanlan.zhihu.com/p/51490163

5.Backward（反向传播）

https://juejin.im/entry/5ac056dc6fb9a028de44d620

https://tigerneil.gitbooks.io/neural-networks-and-deep-learning-zh/content/chapter2.html

https://github.com/INTERMT/BP-Algorithm

https://jdhao.github.io/2016/01/19/back-propagation-in-mlp-explained/

 

图像分割模型：

1. FCN

https://zhuanlan.zhihu.com/p/62839949

http://iblue.tech/2019/05/11/%E8%AE%BA%E6%96%87%E7%AC%94%E8%AE%B0-FCN-%E7%94%A8%E4%BA%8E%E5%9B%BE%E5%83%8F%E5%88%86%E5%89%B2%E7%9A%84%E5%85%A8%E5%8D%B7%E7%A7%AF%E7%BD%91%E7%BB%9C/

https://zh.gluon.ai/chapter_computer-vision/fcn.html

 2.U-Net (E-Net)

https://blog.csdn.net/u012931582/article/details/70215756

https://juejin.im/post/5d63eb7bf265da03e05b2065

https://zhuanlan.zhihu.com/p/31428783

http://tanqingbo.com/2018/10/08/%E8%AE%BA%E6%96%87%E7%AC%94%E8%AE%B0%EF%BC%9A%E7%94%A8%E4%BA%8E%E5%8C%BB%E5%AD%A6%E5%9B%BE%E5%83%8F%E5%88%86%E5%89%B2%E7%9A%84%E5%8D%B7%E7%A7%AF%E7%BD%91%E7%BB%9C/

https://zhuanlan.zhihu.com/p/57530767

 

3. E-Net

https://zhuanlan.zhihu.com/p/39430439

http://hellodfan.com/2018/01/02/%E8%AF%AD%E4%B9%89%E5%88%86%E5%89%B2%E8%AE%BA%E6%96%87-ENet/

https://zhuanlan.zhihu.com/p/31379024

4. Mask-RCNN

https://zhuanlan.zhihu.com/p/37998710

https://www.analyticsvidhya.com/blog/2019/07/computer-vision-implementing-mask-r-cnn-image-segmentation/

https://zhuanlan.zhihu.com/p/40538057

Image Style Transfer（图像风格转变）：

Perceptual Loss： Perceptual Losses for Real-Time Style Transferand Super-Resolution

Feature mimicking：  Mimicking Very Efficient Network for Object Detection

Model distillation： Distilling the Knowledge in a Neural Network

Image Enhancement (图像增强)：

Learning a Deep Single Image Contrast Enhancerfrom Multi-Exposure Images

A Generic Deep Architecture for Single Image Reflection Removaland Image Smoothing  （反射移除）

深度学习框架

caffe教程：

https://blog.csdn.net/m0_38116269/article/details/88119001

https://zhuanlan.zhihu.com/p/24110318

https://absentm.github.io/2016/05/14/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0Caffe%E7%B3%BB%E5%88%97%E6%95%99%E7%A8%8B%E9%9B%86%E5%90%88/