Repository files navigation

In this paper, I propose an architecture of Convolutional Neural Network (CNN) which can jointly learn representations for three tasks: smile detection, gender and age classification.

My model is based on the BKNet architecture. The method used hard parameter sharing, which to reduce the overfit of training separate task.The proposed network takes input from multiple data sources, data were flow through CNN Shared Block which learns joint representations forall tasks from all the sources of data. After the shared block, we separate network into three difference tasks. Each branch then learnstask-specific features and has its own loss calculation method.

python demo.py -i image_path      # estimate imagepython demo.py -v video_path      # estimate videopython demo.py                    # video stream

• python 3.7+
• tensorflow
• numpy
• opencv3.x
• MTCNN for face detection

Firstly, I prepare the training data by merge three datasets. I try to keep the number of training data for each task equally to help have the same impact of each dataset on the model

• IMDB-WIKI Gender Datasets
• IMDB-WIKI Age Datasets
• GENKI-4K Smile Datasets

Runtrainning.ipynb. Change your datasets folder links, training parameters inconfig.py.

Runtesting.ipynb to see result on the test datasets

I also build age-gender-estimation based on efficientnets. If you want, please seeage-gender-estimation

• Effective Deep Multi-source Multi-task Learning Frameworks for SmileDetection, Emotion Recognition and Gender Classification
• Dinh Viet Sang, Le Tran Bao Cuong, Pham Thai Ha, Multi-task learning for smile detection, emotion recognition and gender classification, December 2017