Skip to content

Latest commit

 

History

History
117 lines (79 loc) · 2.8 KB

README.md

File metadata and controls

117 lines (79 loc) · 2.8 KB
Raw

Hopfield Network

Overview

This repository contains an implementation of a Hopfield Network for associative memory tasks, specifically focusing on retrieving images even when they are blurred or corrupted. The network is trained on images generated from the VisualCube API.

Features

  • Associative Memory: Retrieve original images from blurred or noisy versions.
  • Image Processing: Train the network on images and retrive orignal images using blured or lower resolution versions of the image.

Table of Contents

  1. Installation
  2. Usage
  3. License
  4. Contributors
  5. References

Installation

To set up the Hopfield Network project, first install the python packages. Run the following command in your terminal/powershell:

pip install hopfield-memory

or

python -m pip install hopfield-memory

Usage

Import Packages

import hopfield
from matplotlib import pyplot as plt

Initializing the Hopfield Network

hf = HopfieldNetwork(resolution=128*128)

Training and Retrieving Images

#Image URLs / Path / PIL Image Object list
images = ['https://random.com/image.jpg', 'https://random.com/image2.jpg']

# Train the network
hf.train(images, url=True)

# Display a Blurred/Noisy image
blurred_image = 'https://random.com/blurredimage.jpg'
blur = plt.imread(blurred_image)
plt.imshow(blur)
plt.show()

#Retrive the Orignal image from the Blurred/Noisy Image
retrieved_state = hf.retrieve(blurred_image, url=True)
plot_state(retrieved_image, "Retrived Image", pixel=128)

Complete Code

import hopfield
from matplotlib import pyplot as plt

hf = HopfieldNetwork(resolution=128*128)

#Image URLs / Path / PIL Image Object list
images = ['https://random.com/image.jpg', 'https://random.com/image2.jpg']

# Train the network
hf.train(images, url=True)

# Display a Blurred/Noisy image
blurred_image = 'https://random.com/blurredimage.jpg'
blur = plt.imread(blurred_image)
plt.imshow(blur)
plt.show()

#Retrive the Orignal image from the Blurred/Noisy Image
retrieved_state = hf.retrieve(blurred_image, url=True)
plot_state(retrieved_image, "Retrived Image", pixel=128)

For Example Code File Refer To Example.py

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Contributors

  • Vyom Nishant Patel

Acknowledgement

  • Abhisar Mehta
  • Vihaan Mishra

References

- Hopfield, J.J. (1982). "Neural networks and physical systems with emergent collective computational abilities". Proceedings of the National Academy of Sciences. - Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press.

Thanks 😊

Thank you for your patience, and let us know for any suggestions ✨️🤗