【Machine Learning】【Python】二、HoG + PCA + SVM 物体分类 ---- 《SVM物体分类和定位检测》

Jun 14, 2017#Machine Learning120

AI-generated summary

The article discusses the use of HoG, PCA, and SVM for object classification and localization in machine learning using Python. The author initially believed that using PCA to extract top-ranked features would improve classification results, but later realized that it only improved the robustness of the model and reduced false detection rates. The article provides links to the code for HoG, PCA + SVM, and testing SVM. The author also mentions their intention to share PSO code in the future.

** 最新代码 Github 地址： ** https://github.com/HansRen1024/SVM-classification-localization

前言#

当初想当然的以为通过 PCA 降维并提取排名考前的特征可以提高分类效果，结果是我错了。后来也想明白了，因为是同时对正负样本一起降维。也不是一点作用没有，模型的鲁棒性还是有提高的，误检测率有明显下降。

[HoG 代码]#

(https://github.com/HansRen1024/SVM-classification-localization)

[PCA + SVM 代码]#

(https://github.com/HansRen1024/SVM-classification-localization)

[测试 SVM 代码]#

(https://github.com/HansRen1024/SVM-classification-localization)

后记#

等有时间了我上 PSO 代码。