https://github.com/dipanjans/tag-me

The main objective was to design machine learning algorithms for tagging/classifying a collection of images.
https://github.com/dipanjans/tag-me

Last synced: 12 days ago
JSON representation

The main objective was to design machine learning algorithms for tagging/classifying a collection of images.

• Host: GitHub
• URL: https://github.com/dipanjans/tag-me
• Owner: dipanjanS
• Created: 2014-04-11T13:14:53.000Z (over 10 years ago)
• Default Branch: master
• Last Pushed: 2014-04-11T13:58:53.000Z (over 10 years ago)
• Last Synced: 2024-10-07T13:41:07.142Z (about 1 month ago)
• Size: 96.3 MB
• Stars: 2
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
## Tag-Me!

### Main Objective

The main objective was to design machine learning algorithms for tagging/classifying a collection of images.

### Description

Given a collection of images, you need to build a system that can categorize them into one of five 

categories: 'Human Faces', 'Flowers', 'Buildings', 'Cars', or 'Shoes'. This is a multi-class classification 

supervised machine learning problem, where you will have a training set consisting of images and category 

labels and your system has to predict the labels correctly. Ensemble approaches were applied for classification 

and PCA, computer vision and image processing techniques were applied to extract relevant features from images.

Technology used,

 - Python

 - SciPy, NumPy, Mahotas

### Instructions

 1. All the code developed is in the Code folder

 2. Sample features which were extracted by me are in the Sample Extracted Features folder

 3. Results which were obtained are in Results Obtained folder ( same were uploaded by deadline )

 4. Write up is present as a pdf file

 5. Libraries required are in Libraries Installers folder.

For running code,

 - Run the img_processing_utils.py file first to check if there are any errors.

 - Run the submission.py file next to get the results.

Important: These two files should be in the same folder where the "Train", "Validation" and

"Test" folders are present for details look in the code where the path has been fixed.