Faces Vision Demo

iOS11 demo application for age and gender classification of facial images using Vision and CoreML.

Model

This demo is based on the age, gender and emotion neural network classifiers, which were converted from Caffe models to CoreML models using coremltools python package.

Age classification

• Paper
• CoreML model

Gender classification

• Paper
• CoreML model

Emotion Recognition

• Paper
• CoreML model

Requirements

• Xcode 9
• iOS 11

Installation

git clone https://github.com/cocoa-ai/FacesVisionDemo.git
cd FacesVisionDemo
pod install
open Faces.xcworkspace/

Conversion

Download Caffe model and deploy.prototxt. Links can also be found here. Move downloaded files to Covert/EmotionClassification folder.

cd Convert
./download.sh
python age.py
python gender.py
python emotion.py

Download the Age, Gender and Emotion CoreML models and add the files to "Resources" folder in the project's directory.

Build the project and run it on a simulator or a device with iOS 11.

Author

Vadym Markov, markov.vadym@gmail.com

Credits

• Age and Gender Classification using Convolutional Neural Networks
• Emotion Recognition in the Wild via Convolutional Neural Networks and Mapped Binary Patterns

References

• Caffe Model Zoo
• Apple Machine Learning
• Vision Framework
• CoreML Framework
• coremltools