This paper presents an artificial intelligence (AI) system for automatic depression level analysis through visual and vocal expressions. The system uses deep learning techniques to extract features from facial expressions and speech patterns to assess depression severity. Convolutional Neural Networks (CNN) are used for learning deep features, while descriptive raw waveforms are used for visual expressions. The system is designed to detect depression in college students by analyzing their facial features and speech patterns. The system uses image processing techniques for face detection, feature extraction, and classification of these features as depressed or non-depressed. The system is trained with depression features and uses videos of students with frontal faces captured via a webcam to predict depression levels. The system is evaluated using accuracy metrics such as precision, recall, and F1 score. The system also integrates multimodal fusion techniques, combining visual and vocal features for improved depression detection. The system is implemented using Python for the backend and PHP with Bootstrap for the frontend. The hardware requirements include a processor, GPU, RAM, and storage. The software requirements include an operating system, Python, and a database. The system's results show that it achieves good performance on the AVEC2014 dataset, with significant improvements over hand-crafted features. The system's conclusion is that it provides an effective method for automatic depression level analysis through visual and vocal expressions.This paper presents an artificial intelligence (AI) system for automatic depression level analysis through visual and vocal expressions. The system uses deep learning techniques to extract features from facial expressions and speech patterns to assess depression severity. Convolutional Neural Networks (CNN) are used for learning deep features, while descriptive raw waveforms are used for visual expressions. The system is designed to detect depression in college students by analyzing their facial features and speech patterns. The system uses image processing techniques for face detection, feature extraction, and classification of these features as depressed or non-depressed. The system is trained with depression features and uses videos of students with frontal faces captured via a webcam to predict depression levels. The system is evaluated using accuracy metrics such as precision, recall, and F1 score. The system also integrates multimodal fusion techniques, combining visual and vocal features for improved depression detection. The system is implemented using Python for the backend and PHP with Bootstrap for the frontend. The hardware requirements include a processor, GPU, RAM, and storage. The software requirements include an operating system, Python, and a database. The system's results show that it achieves good performance on the AVEC2014 dataset, with significant improvements over hand-crafted features. The system's conclusion is that it provides an effective method for automatic depression level analysis through visual and vocal expressions.