A large-scale face dataset for face parsing, recognition, generation and editing.

A Pytorch implementation face parsing model trained by CelebAMask-HQ, based on EHANet.

Face Parsing via SegNeXt, trained on CelebAMask-HQ

Face Parsing from RGB and Depth Using Cross-Domain Mutual Learning (CVPRW 2021) - IEEE AMFG Acceptance Rate ≈ 27%

TensorFlow implementation of "Context Encoders: Feature Learning by Inpainting" with CelebAMask-HQ Dataset.

Model for automatic portrait segmentation model (aka image matting) to celebrity face(CelebAMask-HQ) dataset.

PyTorch code for binary segmentation on CelebAMask-HQ dataset via both a UNet written from scratch and a pretrained DeepLabv3 model.

Applied the architecture of the BBDM (Brownian Bridge Diffusion Model) in a cross-domain image-to-image translation on CelebAMask-HQ Dataset. Performed the diffusion process in the latent space of the VQGAN.

codes for change CelebAMask-HQ dataset to a set of color segmentation map

Performing Image Segmentation using U-Net to detect facial features like, eyes, lips, ears

No description available

No description available

This repository contains an AI-powered application for face segmentation using BiSeNet model trained on CelebAMask-HQ dataset.

Academic project for Pattern Recognition. The group studied methods for face detection based on skin-color segmentation, using datasets such as CelebAMask-HQ and Pratheepan dataset.

An experimental implementation of MultiRes-U-Net for facial feature segmentation using the CelebAMaskHQ dataset.

celebamask-hq

CelebAMask-HQ

No description available

VAE for image manipulation on CelebAMask-HQ dataset

No description available

Compact ellipse-based facial ROI annotations derived from CelebAMask-HQ skin masks for JPEG-compatible face obfuscation research.

Segmentation of facial attributes according to CelebAMask-HQ annotation.

Semantic segmentation of a subset of CelebAMask-HQ dataset

To create an application which will help in modifying facial features like eyes, ears, nose, etc. 19 total different attributes.

No description available

Lightweight U-Net for CelebAMask-HQ segmentation (~1.7M params)

Face Segmentatjon on CelebAMaskHQ through Deeplabv3

Hybrid UNet face segmentation: encoder_A (CelebAMask-HQ) + decoder_B (LaPa) cross-domain recombination

This project applies the architecture of the Brownian Bridge Diffusion Model (BBDM) to perform cross-domain image-to-image translation on the CelebAMask-HQ Dataset.

Assignment 1 for Advanced Computer Vision [Public Ver.] Experimenting with some xxxNets on CelebAMask facial segmentaion sub-dataset of 1000 [Highest F-score knew so far: 0.8716]