https://github.com/mohamedmetwalli5/image-segmentation

A study of the use of K-means and Normalized cut using KNN on the Berkeley Segmentation Benchmark.
https://github.com/mohamedmetwalli5/image-segmentation

pattern-recognition

Last synced: 7 months ago
JSON representation

A study of the use of K-means and Normalized cut using KNN on the Berkeley Segmentation Benchmark.

• Host: GitHub
• URL: https://github.com/mohamedmetwalli5/image-segmentation
• Owner: MohamedMetwalli5
• License: mit
• Created: 2022-04-05T13:51:57.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2022-05-21T18:21:05.000Z (over 3 years ago)
• Last Synced: 2025-01-08T21:58:15.211Z (9 months ago)
• Topics: pattern-recognition
• Language: Jupyter Notebook
• Homepage:
• Size: 21.9 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
![Language](https://img.shields.io/badge/language-Python%20-blue.svg)

![License](https://img.shields.io/badge/License-MIT%20-red.svg)

# Image Segmentation

## What is this ?

This is a study of the use of K-means and Normalized cut using KNN with k = 5  on the Berkeley Segmentation Benchmark.

## Tunning

The tuning was done on 200 images of the training set to get the best k between {3,5,7,9,11} which is K=3 according to F-measure.

We used the best k to test our algorithm using 50 images of the test set.

## Sample Images

![image](https://user-images.githubusercontent.com/58489322/161772648-17641284-772d-44c0-9d15-4485be20eb52.png)

![image](https://user-images.githubusercontent.com/58489322/161773045-ea346616-9509-42ea-ada5-9a0b48119585.png)

## The Data

To downoad the original used data use this link : http://www.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/BSR/BSR_bsds500.tgz