SENTIMENTAL DATA ANALYSIS USING DEEP LEARNING
DOI:
https://doi.org/10.6084/m9.figshare.26091046Abstract
The project utilizes a cutting-edge convolutional neural network (CNN) model tailored for real-time facial expression recognition. It integrates computer vision methods to interpret emotions from webcam-captured facial images. The system initiates by capturing video frames and applying a face detection algorithm, followed by cropping and resizing the detected faces. These pre processed images undergo classification using a pre-trained CNN model with convolutional, pooling, and fully connected layers. Trained on a diverse facial expression dataset, the model categorizes emotions into seven predefined classes. Predicted emotions are visually represented by displaying corresponding emojis alongside the original images. Additional features include generating reports and statistics detailing usernames, dates, times, and captured images with recognized emojis. TensorFlow, OpenCV, and tkinter libraries are utilized for deep learning, image processing, and the user interface. The project finds applications in user sentiment analysis, market research, and human-computer interaction.