Skip to content

neilchen1998/image-segmentation-project

Repository files navigation

Image Segmentation Project

Image segmentation is a common fashion in computer vision and digital image processing to partition an image into numerous regions and segmentations, based on the naturals of the pixels within the image [1]. The goal of this project is to segment an image of a cheetah from the background using a sundry of approaches.

Overview

We have a black-and-white image of a cheetah. Our goal is to separate the cheetah from the background.

The original image:

cheetah

We use a panoply of algorithms to segment the foreground (cheetah) from the background (grass). Then we compare the results against each other and see which approach yields the best result (the lowest error rate compared with the idea mask).

Dataset [2]

  • cheetah.bmp
  • cheetah_mask.bmp
  • TrainingSamplesDCT_8.mat
  • Zig-Zag Pattern.txt

Approaches

  1. Naive Bayes
  2. Maximum Likelihood (ML)
  3. Maximum a Posteriori (MAP)
  4. Predictive Distribution
  5. Expectation Maximization (EM)

Results

  • Naive Bayes
  • Maximum Likelihood (ML)
  • Maximum a Posteriori (MAP)
  • Expectation Maximization (EM)

Reference

  1. Image Segmentation

  2. UCSD 271A Statistical Learning I

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages