🔗 Dataset link & Raw model link:

• Augmented Dataset (2.4gb)
• Raw model

Requirements :

git clone https://github.com/Ananas1Anass/RealTime-HairColorChange.git
cd RealTime-HairColorChange/
pip install -r requirements.txt

Look into requirements.txt

albumentations==1.3.0
numpy==1.24.1
opencv-python==4.5.4.60
scipy==1.9.1
torch==1.13.1
torchvision==0.14.1
flask==1.1.2
numpy==1.24.1
matplotlib==3.5.2

Running Flask:

Chosen port : 5002 on localhost. If not available, change it to any random port ( except reserved ports )

python inference_web.py

🔗 Papers

• Paper n°1 : Hair detection, segmentation, and hairstyle classification in the wild
• Paper n°2 : Real-time Hair Segmentation and Recoloring on Mobile GPUs

Authors

  author    = {Anass BOUKHEMS,
               Abderrahim ELHAZMI,
               Arthur TODER,
              Jean-Baptiste DEROUET,
              Artem BUZLUKOV,
              Sébastien BATTY,
              Kawtar ABIDINE},
  title     = {Real-time hair color change using Deep Learning}
  year      = {2022}
  emails    =  {boukhemsanass0@gmail.com,
              abderrahimelhazmi@gmail.com,
              toder.arthur@gmail.com,
              derouet.jeanbaptiste@gmail.com,
              artem.buzlukov@gmail.com,
              sebastienbatty@gmail.com,
              Kawtar.abidine00@gmail.com}