20IS76P

Phase 1 Presentation

Facial Emotion Detection

PROJECT GUIDE
Dr. Vidyashree L M.Tech,PhD Team Members
1.Sahana Patil H V
Assistant Professor 2. Chaitra N
3.Reethu M
Department of Information Science and Engineering 4.Venkatesh G
Sri Jayachamarajendra College of Engineering
JSS Science & Technology University, Mysuru
2024-25
Agenda
1. Introduction
2. Problem Statement
3. Aim and Objectives
4. Motivation
5. Challenges
6. Literature Survey
7. Conclusion
Introduction
• Importance of Emotion Recognition: Essential AI application for analyzing facial expressions to identify
emotional states.
• Technological Advances: Enhanced by deep learning, especially Convolutional Neural Networks (CNNs),
improving accuracy and efficiency.
• Project Goals: Develop a model to identify emotions: happiness, sadness, anger, surprise, and neutral
expressions.
• Growing Demand: Increasing need for automated systems that understand human emotions.
• Target Industries: Relevant applications in customer service, mental health, and interactive entertainment.
Problem Statement

Despite the advancements in emotion detection technologies, many existing systems struggle with accuracy and
real-time processing. This project addresses the challenge of accurately recognizing emotions from images,
particularly in varied lighting conditions and diverse facial expressions. The goal is to create a model that can
effectively classify emotions such as happiness, sadness, anger, and surprise.
Aim And Objectives
The primary aim of this project is to create a machine learning model capable of recognizing facial emotions with high accuracy.

The specific objectives include:

1. Develop a Comprehensive System: Create a system that detects emotions from facial expressions.
2. Implement Deep Learning Models: Utilize CNNs for facial analysis
3. Collect and Preprocess Datasets: Gather and prepare diverse emotionally labeled datasets.
4. Achieve High Classification Accuracy: Train models to accurately identify emotions (happiness, sadness, anger, fear, surprise) while
minimizing false positives and negatives.
5. Ensure Real-Time Detection: Optimize the system for quick emotion recognition, suitable for applications like virtual assistants and
customer service bots.
 Motivation
•Enhanced Human-Computer Interaction: Improving interactions between humans and machines by
enabling systems to respond empathetically.

•Improved Customer Service: Facilitating better customer experiences through emotionally responsive AI
systems.

•Mental Health Support: Providing tools for mental health professionals to monitor and assess emotional
states.

•Personalized User Experiences: Tailoring content and responses in entertainment and education based on
users’ emotional states.

•Advancing AI Capabilities: Contributing to the development of more sophisticated AI systems that can
understand and interact with human emotions effectively.
Challenges
•Variability in Expressions: Different cultural backgrounds and individual differences can affect how
emotions are expressed, complicating recognition.

•Occlusions and Lighting: Facial occlusions (e.g., glasses, masks) and varying lighting conditions can
hinder accurate detection.

•Real-Time Processing: Achieving fast and efficient processing in real-time applications is technically
demanding.

•Dataset Limitations: Limited availability of diverse, high-quality, and well-labeled datasets can impact
model training.

•Bias and Ethical Concerns: Potential biases in training data can lead to unfair or inaccurate predictions,
raising ethical issues.
Literature Survey
Sl no Title Author Year Approach Remarks
Development of Real Time Emotion Premananda R, Sunil S. 2024
1 Detection in Faces using Deep Learning Harakannanavar, Sapna
This paper proposes the development of
a Deep Convolutional Neural Network
The study effectively addresses challenges related
to real-time emotion recognition, achieving high
Approach Kumari C, Veena I. (DCNN) model for real-time facial accuracy through deep learning. The results
Puranikmath emotion recognition. The model is indicate its capability for real-world applications,
designed to classify five distinct human with further opportunities for refinement and
emotions by training on datasets such as optimization highlighted..
FER2013 and a local dataset.
Real-Time Emotion Detection using Tanner ,GilliganM.S., Baris 2024
2 Convolutional Neural Networks (CNN) Akis
This paper proposes a Convolutional
Neural Network (CNN) based approach
High Accuracy: The models achieved impressive
accuracy metrics on the test datasets, aligning with
for real-time emotion detection. The state-of-the-art results.Qualitative
authors utilize multiple datasets, Challenges:Despite high quantitative performance,
including the Extended Cohn - Kanade real-time classification revealed limitations,
(CK+) dataset,. particularly with subtle or non-exaggerated
expressions.
A Review of Facial Expression Recognition Scholar Poonam Gujre, 2023
3 using Machine Learning Asst. Prof. Sudha Sharma,
This review paper provides an in-depth
examination of facial expression
The paper is well-structured and thorough,
offering a broad overview of the challenges and
HOD Trapti Sharma, recognition (FER) using machine advancements in the field of facial expression
Director Durgesh Mishra learning, particularly deep neural recognition.
networks.

Emotion Detection Based on Facial -SwapnilSanjay Nagvenkar- 2022

4 Expression Ruchika Prakash Nagekar-
his project focuses on detecting and
classifying emotions from images using
The custom CNN model developed in the study
achieves a higher accuracy (95.80%) compared to
Tanmay Naik- Prof. Amey Convolutional Neural Networks (CNN). VGG16 (87.07%) and ResNet50 (66.32%).
Kerkar

Facial Emotion Recognition Using Deep Wafa Mellouka*, Wahida 2020

5 Learning: Review and Insights Handouzia
This paper presents a comprehensive
review of advancements in facial
The paper effectively synthesizes recent
developments in facial emotion recognition,
emotion recognition (FER) using deep providing valuable insights into the progress made
learning techniques. through deep learning.
Conclusion
The Face Emotion Recognition project successfully demonstrates the application of machine learning techniques
to recognize emotions from facial expressions. Through the systematic approach of setting up the environment,
preparing the dataset, developing and training the model, and finally implementing real-time detection, this
project not only addresses the challenges associated with emotion recognition but also contributes to the growing
field of artificial intelligence. Future work may involve enhancing model accuracy, expanding the dataset, and
exploring more sophisticated neural network architectures.