Skip to content

This repository contains the initial implementation of pre-processing for the VITON dataset, which is the most commonly utilized dataset for virtual try-on tasks.

Notifications You must be signed in to change notification settings

MosbehBarhoumi/VITON-PRE-PROCESSING

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 

Repository files navigation

VITON Pre-Processing Steps

Capture d’écran du 2023-04-26 14-29-17

Introduction

In this project, I have developed a pre-processing pipeline for the VITON dataset. The pipeline includes human-parsing, densepose pose-estimation, and cloth mask. The pre-processing steps are designed for practical uses and are aimed at give you a simple way to pre-process your own dataset the way VITON is processed.

Getting Started

Before you use the notebook, please make a copy of this folder in your drive at this link: https://drive.google.com/drive/folders/1tctFFeiwkLLJymC5rS3RtLRx0YRizRbD?usp=sharing. The folder contains all pre-trained models and code fixed, migrated, and cleaned of errors. Please note that this is the first version and it is not fully automated. I will release other versions in the future.

Usage

The code in the drive is not clean yet, but it works well. If you want to check the results of the pre-processing methods, go to the folder of that pre-processing step and see some examples. The input folder is named 'input-folder' and the output folder is 'output-folder'. To pre-process your own data, just upload it to the input folder and you will get the results in the output folder.

Resources

I would like to mention the resources that I have used for this first try:

Thank you for checking out my project!

About

This repository contains the initial implementation of pre-processing for the VITON dataset, which is the most commonly utilized dataset for virtual try-on tasks.

Topics

Resources

Stars

Watchers

Forks