https://github.com/anubhav-bhargava/logo-classifier

Code for Logo Detection and Classification
https://github.com/anubhav-bhargava/logo-classifier

Last synced: 5 days ago
JSON representation

Code for Logo Detection and Classification

• Host: GitHub
• URL: https://github.com/anubhav-bhargava/logo-classifier
• Owner: Anubhav-Bhargava
• License: mit
• Created: 2018-03-02T15:25:36.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2018-03-04T17:12:40.000Z (over 6 years ago)
• Last Synced: 2024-04-17T12:16:38.560Z (7 months ago)
• Language: Jupyter Notebook
• Homepage:
• Size: 9.48 MB
• Stars: 5
• Watchers: 1
• Forks: 1
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Logo-Classifier

This python code is used to recognize Pepsi company logos in a given image.

# Dataset

It is trained on [Flickr Logos 27 dataset](http://image.ntua.gr/iva/datasets/flickr_logos/)

# Prerequisites

>The code is developed for python 2.7 using a jupyter notebook

>All dependencies are mentioned in Requirements.txt 

# Demo

A gif showing single Pepsi bottle

![Demo](gifs/b.gif)

A gif showing multiple brand bottles

![Demo](gifs/a.gif)

# Algorithms

I've used **Convolution Neural Network** for the classification. 

The model detects Pepsi logo with an accuracy of **~99%**.

It uses Sliding window algorithm(as show above) to detect logo in different parts of the image

# Prediction

To run the prediction on a new image just run the following command

>python evaluate.py --image image_path