Face_Similarity_facial-ratio_and_Embedding-vector_based

Categorical-Time-Series-Prediction-with-Embeddings

DSL 기업연계 프로젝트 (2022. 05)

---

Collaborators

이승재, 황다연, 손예진, 전재현, 이승연, 이승주

---

Requirements

numpy>=1.14.0
pandas>=0.23.4
gdown>=3.10.1
tqdm>=4.30.0
Pillow>=5.2.0
opencv-python>=4.5.5.64
opencv-contrib-python>=4.3.0.36
tensorflow>=1.9.0
keras>=2.2.0
Flask>=1.1.2
mtcnn>=0.1.0
lightgbm>=2.3.1
dlib>=19.20.0
retina-face>=0.0.1
mediapipe>=0.8.7.3
fire>=0.4.0\

File Definition

📁code

#####| Data
#####| Model -- deepface
-- * deepface -- basemodels
  -- commons
  -- weight
  -- preprocessing.ipynb
  -- DeepFace.py
-- get_ratio.py
-- get_ratio_side.py\

        -- penultimate_layer\
        -- Attention\
        -- Tree-based\
        -- Vers\

---

General Process of the Model

• • Crop image to concentrate on the face
• • Get embedding layer for the image
• • Get facial ratio of the image
• • input two variables to the "penultimate layer"
• • Classification task for predicting similarity class

    fron deepface.deepface import DeepFace as DF
    from deepface.deepface import get_ratio, get_ratio_side

    #get embedding vector
    DF.represent( IMG_PATH )

    #get facial ratio
    get_ratio.get_ratio( IMG_PATH )

---

Reference

• https://github.com/serengil/deepface.git