8000 GitHub - jisnoo123/collage: Dataset collager app for images and videos
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

jisnoo123/collage

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

77 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Open Demo in Hugging Face Spaces

Collager

Collager turns your images and videos to dataset collages, with a Gradio UI. The original image is broken into patches and is replace with nearest neighbouring dataset image with respect to the mean pixel value of the three channels. When lookead the result from a distance, one could see the original image (kind of) but replaced with dataset images.

Datasets

  1. Anime
  2. ImageNet10
  3. SVHN
  4. CIFAR_10

Sample

Image

Original Picture of RKMVERI

Sample image

Anime with 50 grids across the height. Zoom into the picture to see the Anime dataset images that replaced your image grids.

Anime

ImageNet10 with 50 grids across the height. Zoom into the picture to see ImageNet10 images.

ImageNet10

SVHN with 50 grids. Zoom into the picture to see SVHN images.

SVHN

CIFAR_10 with 50 grids across the height. Zoom into the picture to see CIFAR_10 images.

CIFAR_10

Video

Original Video

Bleach

ImageNet video with 100 grids across the height.

ImageNet10

Prerequisites

Install the packages required

From working directory, run pip install -r requirements.txt

Creation of processed datasets

In the dataset folder, run create_datasets.py

python3 create_datasets.py

Usage with Gradio UI

Image

In collage, run gradio_app_img.py

python3 gradio_app_img.py

Copy the local server link and paste it in your browser to work in the interface

Image_UI

Video

In collage, run gradio_app_img.py

python3 gradio_app_img.py

Copy the local server link and paste it in your browser to work in the interface

Conclusion

This is my first open source project. I thank my brother Jimut for giving me the idea of this project and Sw. Punyeshwarananda, HOD, Dept. of Computer Science, RKMVERI for providing suggestions for improvements.

Releases

No releases published

Packages

No packages published
0