mood base music recommendation system last yr project

mood based music reccomendation system

1. A Phase II Project Presentation on MOOD BASED MUSIC RECOMMENDATION SYSTEM By Shelke Akshay Sharad Sanap Madhuri Shivaji Rasal Rutuja Devidas Sinare Rajeshree Tulshiram GUIDED BY DR R.S.KHULE Department of Information Technology Matoshri College of Engineering & Research Centre, Nashik . SAVITRIBAI PHULE PUNE UNIVERSITY 2024-2025

2. Agenda Introduction Objectives Literature Review Motivation & Problem Statement Project Requirement specification System Architecture and High level design of the project . Algorithm. Test cases, GUI, Experimental Results , Advantage . Conclusions with future work. References

3. 1. Introduction Music has a profound impact on human emotions, influencing mood and behavio u r in various contexts. In recent years, the integration of emotion recognition technologies with music recommendation systems has gained significant attention, driven by the desire to create more personalized and emotionally responsive experiences. Traditional music players require manual input for song selection, which can be time-consuming and may not always match the user's current emotional state. This project aims to address this gap by developing a music recommendation system that leverages facial emotion recognition to automatically curate playlists based on the user's mood. By employing a Convolutional Neural Network (CNN) for accurate emotion detection, the system captures real-time facial expressions and maps them to specific music genres that correspond to the detected emotions. The ultimate goal is to provide users with a seamless and intuitive music experience that adapts to their emotional needs, enhancing their overall well-being.

4. 2. Objectives To study the literature to understand the challenges in developing and automatic mood recommondation system. To study various techniques to enhance the performance parameters of an automatic mood recommendation system . To develop the mood recommendation system. To Achieve High Accuracy in Emotion Detection . To validate the results of the proposed system.

5. 3.Literature survey NO Title Authors Year Models 1 Facial expression recognition using convolutional neural networks: Pramerdorfer , C., & Kampel , M. 2016 VGG Inception ResNet 2 Facial expression recognition with deep learning Khanzada et al. 2020 ResNet50 SeNet50 VGG-16 Ensemble 3 Deep learning-based facial emotion recognition for human computer interaction applications Chowdary M.K 2021 ResNet50 VGG-19 Inception V3

6. 4.Motivation & Problem Statement The motivation for developing a music recommendation system based on mood detection stems from a few key areas of interest and practical needs: 1.Personalized Listening Experience . 2. Advancement of Artificial Intelligence in Entertainment. 3. Improvement over traditional recommendation system. 4Mental health and emotional support . Problem Statement: The challenge is to design a system that can accurately recommend music based on the real-time detection of a user's mood.

7. 5.Project Requirement specification 5.1 Software Requirements 1 .Operating System: Windows 10/11, macOS, or Linux. 2.Programming Languages: Python 3.6 or higher. 3.IDE/Code Editor: PyCharm, Visual Studio Code, or Jupyter Notebook. 4.Libraries and Frameworks: ○ TensorFlow or Keras for implementing Convolutional Neural Networks. ○ OpenCV for image processing and face detection. ○ NumPy and Pandas for data manipulation.

8. 5 .2Hardware Requirements 1.Processor: Intel Core i5 or equivalent for smooth performance during model training and testing. 3.RAM: Minimum 8GB RAM (16GB recommended) to handle real-time image processing and model predictions. 3.Storage: At least 256GB of SSD for fast read/write operations, especially for handling large datasets. 4.Graphics Card: NVIDIA GPU (e.g., GTX 1060 or higher) if using GPU acceleration for training deep learning models

9. 6.System Architecture and High level design of the project . Fig 6.1: System Architecture 6.1 System Architecture

10. 6.2 Data Flow Diagram Fig 6.2: DF0

11. Start: The process begins. Read Image: The system captures an image, likely of a user's face, using a webcam or similar device. Face Detected?: The system analyzes the image to detect a face. If no face is detected, it loops back to "Read Image . Extract Face Emotion: If a face is detected, the system extracts facial features to determine the user's emotion, such as happiness, sadness, anger, or neutrality. Related Song Found?: The system searches a song database for music matching the detected emotion. If no related song is found, it loops back to "Read Image." Play: If a related song is found, the system plays the music . This system aims to provide personalized music recommendations based on the user's current emotional state. It uses facial expression analysis to understand the user's mood and selects songs accordingly from a database. Detailed DFD diagram

12. 6.3 UML Diagram Fig 6.3: UML Diagram

13. The diagram is a UML sequence diagram illustrating the steps involved in processing an image with facial recognition to detect emotion and play a corresponding sound. The process unfolds as follows: The user initiates the process by starting the system. The user uploads an image containing a face. The system preprocesses the image and detects the face within it. The system analyzes the facial features to detect the emotion. Based on the detected emotion, the system plays a corresponding sound. Sequence diagrams are used to show the interactions between objects in a system in a time-based sequence. They model the high-level interactions between active objects within a system and can model the interactions between object instances within a use case or operation. Detailed UML Diagrams

14. 6.4:Sequance Diagram Fig 6.4: Sequance Diagram

15. 7.Algorithm 7.1 Convolutional Neural Network : A CNN gets a picture as a contribution to the type of a 3D Matrix. The underlying two measurements contrast with the width and height of the image in pixels while the third one identifies with the RGB potential gains of each pixel. CNNs comprises of the accompanying successive modules (every one may contain more than one layer) Convolution : ➢ ReLu activation function ➢ Pooling ➢ Fully connected layers ➢ Output layer Fig.7.1.Basics Convolutional Neural Network

16. Convolution Layer: The part connected with doing the convolution movement in the underlying portion of a Convolutional Layer is known as the Kernel/Channel. Convolution activity is a component savvy network increase activity. Convolutional layers take the three-dimensional information framework and they pass a channel (otherwise called convolutional channel) over the image, applying that to a little window of pixels at the same time (for instance, 3x3 pixels) and this window, being moved until the entire picture has been separated. The convolutional action registers the dab consequence of the pixel regards in the current channel window close by the heaps described in the channel. The yield of this movement is the last tangled picture. The focal point of picture request CNN's is that as the model trains what it really does is that it learns the characteristics for the channel matrices that enable it to remove huge features (shapes, surfaces, concealed districts, etc ) in the image. Each convolutional layer applies one new channel to the tangled image of the past layer that can eliminate one more part. Accordingly, as more channels are stacked, the more features the CNN can remove from an image.The three components that go into the convolution activity are: • Input image • Feature detector • Feature map Fig 7.2.Feature Map generation through convolutional operation

17. Fig 7.3. Creation of Covotional Layer[ ReLu ] Layer: After every convolution activity, CNN applies a Rectified Linear Unit ( ReLu ) function to the yield of the convolved picture. If the convolved image has negative values, it replaces them with ‘0’.It also introduces nonlinearity into the model. 1.Polling Layer 2.Connected Layer 3.Output Layer

18. Pooling Layer: Pooling is the interaction where measurement of the convolved picture is decreased It does as such to diminish handling time and the registering power required. During this cycle, it ensures the fundamental component information. There are a couple of procedures that can be used for pooling. The most generally perceived ones are Max pooling and Typical pooling. In our application, we will use max pooling as it is the best an enormous segment of the events. Max pooling is fundamentally equivalent to the convolution cycle. A window slides over the component guide and thinks tiles of a predefined size. For each tile, max pooling picks the greatest worth and adds it to another component map. In this manner, the face highlights are separated utilizing convolution and pooling layers. Fig 7.4.Maximum pooled Feature Map fully

19. Connected layer: Fully connected layers are a fundamental segment of Convolutional Neural Networks (CNNs), which have been demonstrated fruitful in perceiving and ordering pictures for computer vision. The CNN cycle starts with convolution and pooling, separating the picture into highlights, and investigating them freely. The consequence of this interaction takes care of into a fully connected neural organization structure that drives the last arrangement choice. In the Fully Connected Layer, all neurons of one layer are connected to all neurons in the following layer. Output Layer: The last fully connected layer is the yield layer which applies a SoftMax capacity to the yield of the past fully connected layer and returns a likelihood for each class.

20. 7.2 Detailed study of Model The proposed system contains three modules namely Data Augmentation, Model Training& Testing, Face Detection & Emotion Recognition and Music Recommendation . Fig 7.5Flow Diagram Of System Modules: A. Data Augmentation B. Model Training & Testing C. Face Detection & Emotion Recognition D. Music Recommendation

21. 8.Test Case 8.1 Test Case Table

22. 8.2 Experimental Result Fig 8.2.1 CNN Confusion Matrix Fig 8.2.2CNN Accuracy & Loss plot

23. 8.3 GUI Overview Moodify offers a clean, responsive UI spanning web and mobile: Landing Page : “Welcome to Moodify ” with clear Log In/Get Started buttons. Input Pages : Tabs for text, speech, and camera inputs. Results Page : Displays detected emotion plus recommended songs/Spotify playlist link. Profile & History : Emotion detection log and saved favorites . Settings : Customizable themes (light/dark), Spotify integration, preferences 8.4 Advantages 1.Multi-modal input : Supports text, voice, and facial recognition. 2.Real-time emotion detection : Immediate customized music recommendations. 3.Cross-platform : Available on web (React + Django) and mobile. 4.Spotify integration : Users get live playlists and can save favorites 5.Analytics support : Tracks user emotions over time to refine recommendations.

24. 8.5 SnapShots Fig 8.5.1 Landing Page

25. Fig 8.5.2 Facial Input Page

26. Fig 8.5.3 Result Page

27. Fig 8.5.4 Profile Page

28. 9. Conclusions & Future Work Conclusion: In conclusion, The development of a” Mood-Based Music Recommendation System” represents a significant advancement in personalizing user experiences through the integration of emotional intelligence into technology. By leveraging state-of- theart machine learning algorithms, mood detection models, and recommendation engines, this project demonstrates how technology can adapt to users’ emotional states, offer ing more meaningful and contextually relevant music suggestions. Future Work: Mobile App Launch : Release full-featured Android/iOS version. Privacy First : On-device AI, full control over data Smarter Music Mapping : Use AI + user feedback for personalized songs. Better Emotion Detection : Detect complex emotions like anxiety, etc. Real-Time Tracking : Update playlists based on live mood changes.

29. 10.References 1 Hung, J. C., Lin, K. C., Lai, N. X. (2019). Recognizing learning emotion based on convolutional neural networks and transfer learning. Applied Soft Computing, 84, 105724. https://doi.org/10.1016/j.asoc.2019.105724 2 IFPI. (2018). Global Music Report- Annual state of the industry.Retrieved from IFPI. https://www.ifpi.org/ifpi-global-music-report-2018/[Accessed 15 April 2023]. 3 Bhattarai, B., Lee, J. (2019). Automatic music mood detection using transfer learn ing and multilayer perceptron. International Journal of Fuzzy Logic and Intelligent Systems, 19(2), 88-96. https://doi.org/10.5391/IJFIS.2019.19.2.88 4 Chowdary, M. K., Nguyen, T. N., Hemanth, D. J. (2021). Deep learning-based facial emotion recognition for human computer interaction applications. Neural Computing and Applications, 1-18. https://doi.org/10.1007/s00521-021-06012-8 5 Chung, T. S., Rust, R. T., Wedel, M. (2009). My mobile music: An adaptive personalization system for digital audio players. Marketing Science, 28(1), 5268. https://doi.org/10.1287/mksc.1080.0371

30. Thank You…