Skip to content

EfficientNet - B5 FaceShape Model, Google Mediapipe Facemesh

Notifications You must be signed in to change notification settings

seohyun-kim/hair-consulting-systems

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

74 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

💇‍♀️Hair Consulting Systems

📌 Quick Start

  1. clone
  git clone https://github.com/ESEntial/hair-consulting-demo.git
  cd /hair-consulting-demo
  1. download model file - Google Drive
    save path: /hair-consulting-demo/faceshape_model.pth

  2. execute

 python demo.py

1. Face Shape

1.1. Dataset

Face Shape Dataset - NITEN LAMA


1.2. Model : EfficientNet - B5

reference code: Face Shape - EfficientNetB5 - Acc 68%


1.3. Accuracy : 80%

image


1.4. Predict Test

한소희 박은빈 카리나 김다미 김태리
IMG image 박은빈 카리나 김다미 image
Real Heart Oblong Oval Round Square
Predict Heart Oblong Oval Round Square


2. Facial Features

2.1 Framework

Google Mediapipe - Face Mesh


2.2 Analysis

mid chin philtrum

2.3 Predict Test

face ratio :  4.411731675378876
is_wide_margin =  False
is_long_mid =  False
is_long_chin :  False
is_long_philtrum =  True

3. Hair Style Consult

3.1 Test Report

About

EfficientNet - B5 FaceShape Model, Google Mediapipe Facemesh

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published