Skip to content

Latest commit

 

History

History
89 lines (56 loc) · 3.21 KB

README.md

File metadata and controls

89 lines (56 loc) · 3.21 KB
Raw

Anime Face Generation Using DCGAN

Generated Anime Faces

Table of Contents

Overview

Welcome to the "Anime Face Generation Using DCGAN" project! This repository showcases the application of Deep Convolutional Generative Adversarial Networks (DCGAN) to generate high-quality anime-style faces. DCGANs have demonstrated their ability to create impressive synthetic images, and in this project, we harness this power to craft anime character portraits from scratch.

Dataset

The heart of any machine learning project is the data it's trained on. For this project, we've included the "Anime Faces Dataset" in the data/ directory of this repository. This dataset is a rich collection of anime-style face images, containing a substantial number of samples, each with dimensions [insert dimensions].

Dataset Access: You can find the Anime Faces Dataset in the data/ directory of this repository here.

Prerequisites

To run this project successfully, make sure you have the following prerequisites installed on your system:

  • Python
  • TensorFlow
  • matplotlib
  • numpy

You can quickly install the necessary packages using pip:

How it Works

Our project employs a DCGAN architecture, a type of generative adversarial network specifically designed for image generation tasks. DCGANs consist of two neural networks: a generator and a discriminator. Here's a brief overview of how the system works:

  • Generator: The generator network takes random noise as input and generates synthetic anime face images. Over time, it learns to create more realistic faces.

  • Discriminator: The discriminator network acts as a judge, distinguishing between real anime face images from the dataset and fake images generated by the generator. It provides feedback to the generator on how to improve its generated images.

  • The generator and discriminator are trained simultaneously in a competitive manner. This adversarial training process leads to the generation of increasingly convincing anime faces.

Usage

Training

Follow these steps to train the DCGAN model on your own:

Clone this Repository:

Clone this repository to your local machine using the following command:

git clone https://github.com/umairrrkhan/Anime-Face-Generation-Using-DCGAN.git

Navigate to the Project Directory:

Move into the project directory:

cd Anime-Face-Generation-Using-DCGAN.ipynb

You can monitor the training progress and generated images using TensorBoard or other visualization tools.

Generating Anime Faces

After training, you can generate anime faces using the trained model:

bash Copy code python generate.py Results Here's a glimpse of some anime faces generated by our DCGAN model:

Generated Anime Faces

Contact

If you have any questions, feedback, or inquiries about this project, don't hesitate to get in touch with us at [email protected] .