سنة النشر

2024

العنوان

Deep Learning and Histogram Gradient Algorithm to Detect Visual Information Based on Artificial Intelligent

الخلاصة

The realm of computer vision and machine learning is of paramount importance in comprehending human emotions and has garnered considerable interest in the field of face emotion recognition.to recognize the human expression from video is considered a difficult and interesting task since the face is the main means of communication and the most communicative part of the body to display emotions. In this paper provides a detailed investigation of facial expression detection and its application to a proposed real voice expression to make a comparison system that relies on emotions of human, by using deep learning algorithm to recognize and classify emotions shown in face photographs. By using (CNN) (one dimension & two dimensions), with layers and time distributed in order to lead to the efficient capturing of emotion from the video sequences , with system can doing train and testing to the dataset from CREMA -D dataset. The developing a new architecture for video-based emotion recognition. Preprocessing, feature extraction, and classification utilizing 1D-CNN and 2D-CNN models comprise the architectural framework. The 1D-CNN model classifies features after being extracted by the main algorithm of extraction of (HOG), while the 2D-CNN model features extraction and classification simultaneously. The accuracy of the CNN 1D model is 0.81 , which indicates that it produces high results. Additionally, the 2D CNN worked superbly, with an accuracy score of 0.68. © 2024 ACM.