- Mousavi, S.M.H., Ilanloo, A. (2025). Bees Local Phase Quantisation Feature Selection for RGB-D Facial Expression Recognition. In: Pham, D.T., Hartono, N. (eds) Intelligent Engineering Optimisation with the Bees Algorithm. Springer Series in Advanced Manufacturing. Springer, Cham. https://doi.org/10.1007/978-3-031-64936-3_12
This repository implements Bees Local Phase Quantization (LPQ) Feature Selection for RGB-D Facial Expressions Recognition, as outlined in our paper. We used the Bees Algorithm (BA) to optimize LPQ-extracted features for classifying RGB-D images of facial expressions (emotion recognition), using SVM, KNN, Shallow Neural Network, and Ensemble Subspace KNN classifiers.
This project uses the Iranian Kinect Face DataBase (IKFDB), containing RGB and depth images for five emotions. Our pipeline includes preprocessing, LPQ feature extraction, Bees Algorithm (BA) feature selection, and classification. Key findings show that BA significantly boosts accuracy, especially with the Ensemble Subspace KNN classifier, reaching up to 99.8%.
This book chapter, "Bees Local Phase Quantisation Feature Selection for RGB-D Facial Expression Recognition," featured in the book "Intelligent Engineering Optimisation with the Bees Algorithm," part of the Springer Series in Advanced Manufacturing.This work was initially presented at "The International Workshop on the Bees Algorithm and its Applications" in 2022, hosted by the University of Birmingham.
