Facial Expression Recognition by Regional Weighting with Approximated Q-Learning

Application of artificial intelligence methods for monsoonal river classification in Selangor river basin, Malaysia

Rivers in Malaysia are classified based on water quality index (WQI) that comprises of six parameters, namely, ammoniacal nitrogen (AN), biochemical oxygen demand (BOD), chemical oxygen demand (COD), dissolved oxygen (DO), pH, and suspended solids (SS). Due to its tropical climate, the impact of seasonal monsoons on river quality is significant, with the increased occurrence of extreme precipitation events; however, there has been little discussion on the application of artificial intelligence models for monsoonal river classification. In light of these, this study had applied artificial neural network (ANN) and support vector machine (SVM) models for monsoonal (dry and wet seasons) river classification using three of the water quality parameters to minimise the cost of river monitoring and associated errors in WQI computation. A structured trial-and-error approach was applied on input parameter selection and hyperparameter optimisation for both models. Accuracy, sensitivity, and precision were selected as the performance criteria. For dry season, BOD-DO-pH was selected as the optimum input combination by both ANN and SVM models, with testing accuracy of 88.7% and 82.1%, respectively. As for wet season, the optimum input combinations of ANN and SVM models were BOD-pH-SS and BOD-DO-pH with testing accuracy of 89.5% and 88.0%, respectively. As a result, both optimised ANN and SVM models have proven their prediction capacities for river classification, which may be deployed as effective and reliable tools in tropical regions. Notably, better learning and higher capacity of the ANN model for dataset characteristics extraction generated better predictability and generalisability than SVM model under imbalanced dataset.

Feature Selection on 2D and 3D Geometric Features to Improve Facial Expression Recognition

An essential aspect in the interaction between people and computers is the recognition of facial expressions. A key issue in this process is to select relevant features to classify facial expressions accurately. This study examines the selection of optimal geometric features to classify six basic facial expressions: happiness, sadness, surprise, fear, anger, and disgust. Inspired by the Facial Action Coding System (FACS) and the Moving Picture Experts Group 4th standard (MPEG-4), an initial set of 89 features was proposed. These features are normalized distances and angles in 2D and 3D computed from 22 facial landmarks. To select a minimum set of features with the maximum classification accuracy, two selection methods and four classifiers were tested. The first selection method, principal component analysis (PCA), obtained 39 features. The second selection method, a genetic algorithm (GA), obtained 47 features. The experiments ran on the Bosphorus and UIVBFED data sets with 86.62% and 93.92% median accuracy, respectively. Our main finding is that the reduced feature set obtained by the GA is the smallest in comparison with other methods of comparable accuracy. This has implications in reducing the time of recognition.