Abstract: This article addresses the use of extraction of characteristics for facial expressions in conjunction along a neural network to identify various sentiments (happy, sad, angry, fear, surprised, neutral etc..). The rapid advancement of artificial intelligence has made significant contributions to the field of technology. One of the most crucial aspects of human communication that aids in our understanding of what the other person is attempting to say is the expression on their face. Only one-third of the message is understood vocally, and the other two-thirds are understood nonverbally. There are numerous face emotion recognition (FER) systems in use today, however they are ineffective in real-world situations. Despite the fact that many assert that their system is nearly perfect. The testing data then examines the information and its classification report names the testing data and indicates how accurately it was classified. For better data categorization, many strategies are used, modifying the images using the Histogram of Oriented Gradients (HOG) and Discrete Wavelet Transform or passing the training images through a Gabor filter (DWT).The training images are first run through The Histogram of Oriented Gradients (HOG) produces the best results to date, with an average precision of 92%.

Keywords: Facial Expressions, Face Emotion Recognition (FER), Histogram of Oriented Gradients (HOG), Discrete Wavelet Transform (DWT).

PDF | DOI: 10.17148/IJARCCE.2023.12220

Open chat
Chat with IJARCCE