Home

Awesome

ACGAN Conditional Anime Generation

Note: I have restructured the whole folder, cleaned up training code, pruned my dataset, and updated most of the results. You can see the old version of this repo in old.

Generate colorful anime characters using GAN. <br>

<p align="center"> <img src="results/change_hair.png" /> </p> <p align="center"> <img src="results/change_eye.png" /> </p>

Generating characters with different hair and eye colors

By interpolating generator input, we can see some interesting results.

<p align="center"> <img src="results/interpolate.png" /> </p>

By fixing color classes and varying noise, we can generate anime characters with same colors but different identity.

<p align="center"> <img src="results/blue_eye_red_hair.png" /> </p> <p align="center"> <img src="results/green_eye_purple_hair.png" /> </p> <p align="center"> <img src="results/orange_eye_blue_hair.png" /> </p> <p align="center"> <img src="results/purple_eye_pink_hair.png" /> </p> <p align="center"> <img src="results/red_eye_black_hair.png" /> </p>

Generate characters with different artistic styles

I noticed that anime produced in different years have distinctive artistic styles. To name a few:

By conditioning the ACGAN on year labels, we can generate characters with different artistic styles.

<p align="center"> <img src="results/year_selected.png" /> </p>

Interpolating along the year latent code:

<p align="center"> <img src="results/year_morph.png" /> </p>

Start training

Modify config.yaml as you wish.

> python3 run.py

The dataset

The current dataset is a composition of 2 datasets:

  1. One dataset with eye and hair color labels (30k+ images)
  2. One dataset with year label, which is the year in which the anime is produced (60k+ images).

The dataset format is as follows:

- images/
    - XXXXX.jpg
    - ...
- labels.pkl
- eye_label.json
- year_label.json
- hair_label.json

After loading in labels.pkl with pickle, you will get a dictionary of { filname : labels }. The labels are formatted as (eye, hair, year) tuples.

{
    "32455": (8, 10, 5),
    ...
}

This means 32455.jpg has eye class 8, hair class 10, year class 5.

Missing labels will be a None. All images from dataset 1 will have year labels None, while all images from dataset 2 will have eye and hair label None.

Source code in the current repo is used to train on the first dataset. This requires some manual preprocessing (see dataset/anime_dataset.py) to extract the first dataset from the whole dataset.

The .json files map discrete labels to semantics.

// eye_label.json
{
    "aqua": 0, 
    "black": 1, 
    ...
}

Some notes