Real-Time Speech Emotion Recognition Using a Pre-trained Voice Classification Network

Project Title

Real-Time Speech Emotion Recognition Using a Pre-trained Voice Classification Network

Project Area of Specialization Software EngineeringProject Summary

Speech Emotion Recognition is the task of recognizing the emotional aspects of speech irrespective of the semantic contents. While humans can efficiently perform this task as a natural part of speech communication, the ability to conduct it automatically. The idea behind creating this project was to build a machine learning model that could detect emotions from the speech we have with each other all the time. Nowadays personalization is something that is needed in all the things we experience everyday. So why not have a emotion detector that will guage your emotions and in the future recommend you different things based on your mood. This can be used by multiple industries to offer different services like marketing company suggesting you to buy products based on your emotions, automotive industry can detect the persons emotions and adjust the speed of autonomous cars as required to avoid any collisions etc.

Project Objectives

Even though it isn't that popular, SER has entered so many areas these years, including:

1. The medical field: In the world of telemedicine where patients are evaluated over mobile platforms, the ability for a medical professional to discern what the patient is actually feeling can be useful in the healing process.
2. Customer service: In call center conversation may be used to analyze behavioral study of call attendants with the customers which helps to improve the quality of service.
3. Recommender systems: Can be useful to recommend products to customers based on their emotion towards that product.

Project Implementation Method

First, we gonna need to install some dependencies using pip:

1. Librosa
2. Numpy
3. Soundfile
4. Scikit-learn
5. PyAudio

The whole pipeline is as follows (as same as any machine learning pipeline):

1. Preparing the Dataset: Here, we download and convert the dataset to be suited for extraction.
2. Loading the Dataset: This process is about loading the dataset in Python which involves extracting audio features, such as obtaining different features such as power, pitch and vocal tract configuration from the speech signal, we will use librosa library to do that.
3. Training the Model: After we prepare and load the dataset, we simply train it on a suited sklearn model.
4. Testing the Model: Measuring how good our model is doing.

Benefits of the Project

1. Emotion plays a significant role in daily interpersonal human interactions. This is essential to our rational as well as intelligent decisions.
2. It helps us to match and understand the feelings of others by conveying our feelings and giving feedback to others.
3. Several inherent advantages make speech signals a good source for affective computing. For example, compared to many other biological signals (e.g., electrocardiogram), speech signals usually can be acquired more readily and economically. This is why the majority of researchers are interested in speech emotion recognition.
4.  SER aims to recognize the underlying emotional state of a speaker from her voice.

Technical Details of Final Deliverable

1. Librosa
2. Numpy
3. Soundfile
4. Scikit-learn
5. PyAudio

Final Deliverable of the Project Software SystemCore Industry ITOther IndustriesCore Technology OthersOther TechnologiesSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	40000
TOOLS	Equipment	4	10000	40000