Creation of an application for real time emotions recognition on a low resource machinies

Authors

  • Донецький національний університет імені Василя Стуса
  • І.В. Богач Вінницький національний технічний університет

Abstract

The improvements related to construction of better algorithms and big data availability gave a possibility to solve a huge number of vision related tasks. One of such problems is understanding and detection of facial expressions. This task is crucial for many applications including market research, making self-driving cars safer, analysis of interviews, etc. The described task is not a new one, thus many solutions exist along with datasets, including FER [1] and AffectNet [2]. However, the task becomes more challenging if we consider detecting a new class of images, the balance between accuracy, latency and size of the model, hyperparameters tuning. The aim of this work is to give exhaustive answers to all the aforementioned questions via description of a successful application serving a neural network trained on a custom dataset for emotion detection.

Author Biographies

, Донецький національний університет імені Василя Стуса

студент 4 курсу

І.В. Богач, Вінницький національний технічний університет

к.т.н., доцент

References

Ali Mollahosseini, Behzad Hasani, and Mohammad H. Mahoor, “AffectNet: A New Database for Facial Expression, Valence, and Arousal Computation in the Wild”, IEEE Transactions on Affective Computing, 2017.

I. J. Goodfellow, D. Erhan, P. L. Carrier, A. Courville, M. Mirza, B. Hamner, W. Cukierski, Y. Tang, D. Thaler, D.-H. Lee, Y. Zhou, C. Ramaiah, F. Feng, R. Li, X. Wang, D. Athanasakis, J. Shawe-Taylor, M. Milakov, J. Park, R. Ionescu, M. Popescu, C. Grozea, J. Bergstra, J. Xie, L. Romaszko, B. Xu, Z. Chuang, and Y. Bengio. Challenges in representation learning: A report on three machine learning contests. Neural Networks, 64:59--63, 2015. Special Issue on "Deep Learning of Representations"

Henrique Siqueira, Sven Magg and Stefan Wermter. Efficient Facial Feature Learning with Wide Ensemble-based Convolutional Neural Networks. [Electronic resource] – Electronic data. – Mode of access: https://arxiv.org/pdf/1612.02903.pdf – Title from the screen.

Charlie Hewitt, Hatice Gunes. CNN-based Facial Affect Analysis on Mobile Devices. [Electronic resource] – Electronic data. – Mode of access: https://arxiv.org/pdf/1807.08775.pdf – Title from the screen.

Oahega - Emotion detector. [Electronic resource] – Electronic data. – Mode of access: https://play.google.com/store/apps/details?id=org.oahega.com – Title from the screen.

Downloads

Published

2022-06-30

Issue

Section

Секція "Системи та методи аналізу даних та підтримки прийняття рішень"