<p>Emotion recognition (ER) plays a crucial role in enhancing human-machine interaction, offering insights into the mental state of individuals based on physiological and behavioral cues. This paper proposes a deep learning approach using a Convolutional Neural Network (CNN) optimized with Layer-wise Adaptive Moments (LAMB) for the classification of emotions utilizing physiological signals data. The model classifies four core emotions—Angry, Happy, Neutral, and Sad—detected from emotions via a ConvNet architecture, while incorporating heart rate (in beats per minute) and temperature data obtained from pulse, temperature, and electrocardiogram (ECG) sensors. By combining expression recognition with physiological data, the system achieves a holistic understanding of emotional states. A comparative analysis between the proposed CNN model and traditional classifiers demonstrates the efficacy of the deep learning (DL) approach, by reaching an accuracy for four emotions such that, Angry (Class-0) is 98.41%, Happy (Class-1) is 98.32%, Sad (Class-2) is 98.43%, and Relaxed (Class-3) is 98.25% on a dataset of 115 samples. This integrated solution not only improves emotion detection accuracy but also serves as a potential tool for alerting individuals to prolonged negative emotional states, encouraging timely intervention or consultation.</p>