MOG

Explores fitting parametric models MOG to visual data, and perform inference with the model.

Requirment

Tested against Python 3.6.6. Required Python modules:

• numpy
• scipy
• matplotlib
• pillow

---

Introduction

Part I

practicalMixGaussA.ipynb fits one Gaussian model to the data for skin and another Gaussian to non-skin pixels, and use this to find the posterior probability that each pixel in an image is skin.

practicalMixGaussB.ipynb fits a mixture of Gaussians model to one dimensional data.

practicalMixGaussC.ipynb fits a mixture of Gaussians model to the RGB data.\

Part II

practicalMixGauss_Apples.ipynb trains a mixture of Gaussians model and use it to inferece apples in unknown image.