Ensemble-based face expression recognition approach for image sentiment analysis

<span>Sentiment analysis based on images is an evolving area of study. Developing a reliable facial expression recognition (FER) device remains difficult challenge as recognizing emotional feelings reflected in image dependent diverse set factors. This paper presented ensemble-based model for FER that incorporates multiple classification models: i) customized convolutional neural network (CNN), ii) ResNet50, and iii) InceptionV3. The averaging ensemble classifier method used to the predictions from three models. Subsequently, proposed trained tested dataset with uncontrolled environment (FER-2013 dataset). experiment demonstrated ensembling classifiers outperformed all single classifying positive neutral expressions (91.7%, 81.7% 76.5% accuracy rate happy, surprise, neutral, respectively). However, when disgust, anger, sadness, ResNet50 alone better choice. Although Custom CNN performs best fear (55.7% accuracy), can still classify comparable performance (52.8% accuracy). potential using enhance FER. As result, has shown 72.3% rate.</span>