Oppositional Brain Storm Optimization With Deep Learning Based Facial Emotion Recognition for Autonomous Intelligent Systems

Autonomous Intelligent Systems (AIS) states to a class of intelligent devices that manage and create decisions independently without human interference. These techniques use numerous models contains artificial intelligence (AI), robotics, machine learning (ML), and sensor fusion, in order to identify its environment, reflect that data, and execute accordingly to achieve specific goals. Facial emotion detection in AIS applies to the capability of AI-driven autonomous machines to detect and interpret human emotions reliant on facial expressions. This technology permits AIS to identify and reply to the emotional conditions of individuals they interconnect with, foremost to more normal and empathetic human-machine communications. So, this study develops an Oppositional Brain Storm Optimizer with Deep Learning based Facial Emotion Recognition (OBSODL-FER) system for AIS. The foremost goal of the OBSODL-FER system is to identify and organize dissimilar classes of facial emotions of the drives in autonomous vehicles. To achieve, the OBSODL-FER approach mainly employs an Xception-based deep convolutional neural networks (CNNs) for feature extractor. Also, the developed OBSODL-FER approach exploits the OBSO system for the hyperparameter selection of the Xception method. Besides, an improved LSTM model (ILSTM) is applied to the classification procedure. Furthermore, a jellyfish search (JFS) optimizer is employed for the optimum hyperparameter selection of the ILSTM technique. The simulation results of the OBSODL-FER approach are verified on a benchmark facial emotion dataset. The experimental results inferred the enhancement of the OBSODL-FER system over other DL algorithms.