Year: 2026 | Month: April | Volume: 13 | Issue: 4 | Pages: 278-284
DOI: https://doi.org/10.52403/ijrr.20260427
Emotion Recognition and Content Recommendation System Using Deep Learning
Muppidi Sai Kumari1, Manogna Nannapaneni1, Pothula Gayathri2, Reddybattula Jahnavi3, Janjanam Jasmitha4, Midatani Manoj5
1,2,3,4,5Department of Information Technology,
Dhanekula Institute of Engineering & Technology, JNTUK, Vijayawada, Andhra Pradesh, India.
Corresponding Author: Manogna Nannapaneni
ABSTRACT
Facial Emotion Recognition (FER) has become Emotion Recognition in the Context of Human Computer Interaction and Affective Computing. An important component in intelligent human computer interaction and affective computing systems. An integrated approach to Emotion Recognition is needed. Deep learning-based emotion recognition system is developed for affect detection and classification model for facial images with the goal of enabling secure interaction and providing users with personalized suggestions. The affect detection and classification model for facial images uses a pre-trained MobileNetV2 architecture which is fine-tuned for the classification of seven
emotional states: angry, disgust, fear, happy, neutral, sad, and surprise. We apply pre-processing and image augmentation techniques on the input images to improve the model generalization performance and robustness. We then proceed to optimize using the Adam optimizer with cross-entropy loss and label smoothing. At inference time, this model generates emotion predictions along with confidence scores and applies time-aware bias adjustment to improve contextual relevance. Recommendations tailored to the current emotional state of the user are obtained via a secure login-enabled interface based on the emotions identified. Experimental validation of the proposed approach is also provided. accuracy of 97.89%, achieving high precision and recall across all classes with minimal overfitting. The lightweight architecture and integrated system design enable efficient real-time deployment in user-centric applications.
Keywords: Facial Emotion Recognition, Deep Learning, Transfer Learning, MobileNetV2, Affective Computing.
[PDF Full Text]