AURA PULSE

Supervisor: Presented By:

Dr. Monika Nagar Shivang Rastogi
Assistant Professor Prince Kumar
Department of Computer Science and Design Nensi Varshney
IMSEC, Ghaziabad Vivek Chaudhary
Introduction
Did you know that over 1 in 4 people globally face mental health challenges, yet the majority
never seek help?
Rising levels of stress, anxiety, and burnout highlight the urgent need for effective, tech-enabled
emotional support systems.

Aura Pulse addresses this need by leveraging advanced AI and computer vision to detect emotions
through facial expressions.
By analyzing facial cues in real time via a webcam, Aura Pulse provides personalized music and
lifestyle recommendations tailored to the user's current emotional state.

Music has a profound connection to human emotions. Aura Pulse uses this link to uplift or soothe the
user’s mood, offering suggestions that support emotional well-being.
Aura Pulse is a seamless blend of technology and empathy—a reliable, real-time companion for
mental wellness.
Problem statement
Despite advancements in AI and emotion recognition, several key challenges remain in developing a
personalized, emotion-aware music and wellness recommendation system:
• Emotion Detection :
Achieving reliable and precise emotion recognition from facial expressions in varied lighting and
user conditions.
• Real-Time Processing:
Ensuring low-latency analysis for real-time emotion detection and immediate response.
• Personalized Recommendation:
Generating accurate music and lifestyle suggestions tailored to individual emotional states.
• System Integration:
Seamlessly integrating emotion detection with recommendation engines and user interfaces.
• Dataset Diversity:
Training models on diverse facial expression datasets to improve generalization across users.
• User Experience & Acceptance:
Designing an intuitive, non-intrusive system that encourages regular use and emotional
engagement.
Objective
• Emotion Detection
Develop a system capable of accurately detecting real-time emotions using facial recognition
techniques.
• Personalized Recommendations
Recommend music and lifestyle suggestions aligned with the user's current emotional state.
Design an intelligent music player that adapts to user behavior and listening patterns for enhanced
personalization.
• User Experience & Well-being
Create a user-friendly, engaging interface that encourages emotional expression and daily use.
Promote emotional well-being by offering mood-enhancing suggestions and wellness tips based
on detected emotions.
• Adaptive Learning
Utilize machine learning to continuously learn from user interactions and feedback, refining
recommendations over time.
Literature Survey
Author(s) Year Contribution

Integrated emotion recognition with intelligent devices;

Guo et al. 2024 highlighted challenges in dynamic, real-time emotion
detection.

Achieved 80% accuracy in emotion-based music

Kranthi Kiran et al. 2024 recommendation using facial recognition, showing
potential for mental healthcare.

Used CNNs to detect facial expressions and generate

Padmavati et al. 2024
personalized playlists for music therapy.

Combined facial expression recognition and music

Rajjan et al. 2023 recommendation; focused on detecting micro-expressions
with 62.1% accuracy.
Literature Survey
Author(s) Year Contribution

Real-time emotion detection using CNNs;

Athavle et al 2021 utilized Bollywood song database to improve
mood alignment.

Employed GANs for emotion-aware music

Bhaumik et al. 2021 recommendations, enhancing accuracy and
personalization.

Developed a real-time facial expression-based

Florence et al. 2020 music recommendation system with over 80%
emotion detection accuracy.
Methodology used
1. Data Collection & Preprocessing
• Collected facial emotion datasets from open-source repositories.
• Performed data cleaning, noise reduction, and balancing to ensure high-quality model training.
2. Emotion Detection using CNN
• Trained Convolutional Neural Networks (CNNs) to identify user emotions from facial
expressions in real time.
• Fine-tuned the model to improve detection accuracy across diverse faces and lighting conditions.
3. Recommendation System
• Implemented a hybrid approach combining collaborative filtering and content-based algorithms.
• Generated personalized music and lifestyle suggestions based on the detected emotion.
4. User Testing & Feedback Loop
• Conducted beta testing with real users to evaluate accuracy, usability, and satisfaction.
• Integrated user feedback for iterative improvements to the model and interface.
5. Deployment & Scalability
• Developed a scalable infrastructure compatible with web and mobile platforms.
• Ensured smooth performance across devices with responsive design and efficient resource usage.
System Overview
Flow Chart
Requirement Analysis
1. Software Requirements:
• Python, JavaScript
• TensorFlow, OpenCV
• MySQL
2. Hardware Requirements:
• Standard computer or mobile device
• Integrated or external webcam
3. Functional Requirements:
• Real-time emotion detection
• Emotion-based music and lifestyle recommendations
• User interface for interaction and feedback
• Data storage and management
• Efficient system performance across platforms
Conclusion
1. Innovative Solution
Aura Pulse leverages advanced facial emotion recognition to deliver personalized support,
combining AI with emotional intelligence.
2. Real-Time Emotion Detection
The system accurately identifies user emotions in real time, enabling timely and relevant responses.
3. Personalized Recommendations
By offering emotion-based music and lifestyle suggestions, Aura Pulse enhances user engagement
and supports emotional well-being.
4. Broader Impact
Beyond entertainment, Aura Pulse holds significant potential in mental health support, acting as a
non-invasive, tech-driven wellness companion.
5. Empowering Users
The platform encourages self-awareness and emotional expression, promoting proactive mental
health management.
Future Scope
1. Multimodal Emotion Recognition
Integrate additional biometric signals such as heart rate or galvanic skin response with facial
expression analysis for deeper emotional insight.
2. Personalized Emotion Models
Train emotion recognition systems using individual user data to tailor recommendations based on
unique emotional responses and preferences.
3. Culturally Adaptive AI
Incorporate cultural and regional variations in facial expressions to make the system more inclusive
and globally effective.
4. Context-Aware Recommendation
Enhance personalization by factoring in contextual cues like time of day, location, or user activity
during emotion analysis.
5. Explainable AI
Implement interpretable models that allow users to understand how their emotions were detected
and how recommendations were generated.
6. Voice-Based Emotion Detection
(Optional/For Later) Extend the system to analyze vocal tone and speech patterns to support
multimodal emotional input.
References
1. Guo, R., Guo, H., Wang, L., Chen, M., Yang, D., & Li, B. (2024). Development and application of
emotion recognition technology – A systematic literature review. BMC Psychology.
2. Kiran, B. K., Shanthan, P., Ram, K. S., & Usha, K. (2024). Emotion-based music recommendation
system using VGG16-CNN architecture. International Journal for Research in Applied Science &
Engineering Technology (IJRASET), 12(6).
3. Padmavati, K., Kumar, D. A., Ganavi, S. M., Keerthi, K. C., & Sourabha, Y. N. (2024).
SoulSound: Enhancing musical therapy through facial expression recognition with machine learning.
International Research Journal of Engineering and Technology (IRJET).
4. Rajjan, M., Deore, P., Mohite, Y., & Desai, Y. (2023). Harmonic Fusion: AI-driven music
personalization via emotion-enhanced facial expression recognition. International Journal of
Innovative Science and Research Technology, 8(12).
5. Athavle, M., Mudale, D., Shrivastav, U., & Gupta, M. (2021). Music recommendation based on
face emotion recognition. Journal of Informatics Electrical and Electronics Engineering, 2(2).
6. Bhaumik, M., Attah, P. U., & Javed, F. (2021). Emotion-integrated music recommendation system
using generative adversarial networks. SMU Data Science Review, 5(3), Article 4.
7. Florence, S. M., & Uma, M. (2020). Emotional detection and music recommendation system based
on user facial expression. IOP Conference Series: Materials Science and Engineering, 912(6),
062007.