# image-pca

**Repository Path**: williamzjc/image-pca

## Basic Information

- **Project Name**: image-pca
- **Description**: image pca (仅供毕业设计)
- **Primary Language**: Python
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2019-11-13
- **Last Updated**: 2021-01-09

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# image-pca

PCA for image

[PCA原理](/PCAcn.md)

**仅供毕业设计！！！**

## 介绍
只是一个PCA wrapper (scikit-learn). fit 等方法的输入式图片列表而不是数组. 还可直接用 ezfit.

主程序：impca.py

辅助：PIL_ext.py（为扩展Pillow而写）

## 依赖
- numpy
- scikit-learn
- pillow
- PIL_ext (written by the author)

## 使用

```python
    # construct PCA model with 2 components
    ip = ImagePCA(n_components=2)
    # images = a list of image (based on pillow)
    ip.fit(images)
    # ip.ezfit(folder) # eazy version of fit method, folder is the folder of images
    # get and save eigen images
    for k, im in enumerate(ip.eigen_images):
        im.save(f'eigen{k}.jpg')
    # get coordinates and errors of images (that could be new)
    coords = ip.coordinate(images)
    errors = ip.error(images)
    
    # see demo function
```

## 实验
```python
from impca import *
demo(n_components=2)  # 请设置图片文件夹路径
```
见两张图片 (eigen images) 和 .csv 文件.

![eigen 0](eigen0.jpg)
![eigen 1](eigen1.jpg)