Skip to content

SongChaeYoung98/real-time-face-de-identification

Repository files navigation

real-time-face-de-identification

De-identification of a Human's Face in Real-Time with tensorflow and keras

Project Structure

real-time-face-de-identification
├── datasets
│   ├── etc
│   ├── face_dataset_test_images
│   │   ├── me
│   │   └── not_me
│   ├── face_dataset_train_aug_images
│   │   ├── me
│   │   └── not_me
│   └── face_dataset_train_images
│       ├── me
│       └── not_me
├── models
├── All Installed Package Versions.txt
├── README.md
├── classifier.py : Build Your Own Classifier
├── classifier_model_comparison.py : Build 5 types of models and compare
└── real-time-de-identification.py : De-identify a specific person's face in real time

Datasets

Installation

Tensorflow_gpu-2.5.0

  • CUDA 11.2
  • cuDNN 8.1

Requirement

  • keras
  • opencv-python
  • matplotlib
  • numpy>=1.20.3
  • scipy
  • seaborn
  • tensorflow(gpu)>=2.5.0

Models

  • VGG16
  • ResNet50
  • ResNet152
  • Xception
  • MobileNet
  • Custumized Classifier

Language

  • Python 3.8

Development Environment

  • PyCharm 2020.2.3
  • VM: Anaconda3
  • CPU: Intel(R) Core(TM) i7-7700
  • GPU: NVIDIA GeForce GTX 1060 6GB
  • OS: Window 10 x64
  • RAM: 16GB

Releases

No releases published

Packages

No packages published

Languages