Emotion Recognition AI

Student Information

Student Name:

Hanson Liew

Course:

Managing IT Projects

Semester:

Fall 2025

📋 1. Project Title & Description

Project Title: Emotion Recognition AI

Build an AI model for recognizing facial expressions and determining emotional responses. This model will take webcam data and, in real time, analyze the emotion of a person captured by the camera. I chose this AI-focused project because I haven’t built an AI model before, and I believe it’s a crucial skill for the future.

🎯 2. Vision, Mission & Strategy Alignment

Personal Mission Connection:

This project helps me achieve a developing niche in the field of Cybersecurity that I can leverage to become a leading figure in the industry.

Vision Alignment:

This project shows my proficiency in an emerging tool and can help me launch my career by helping me stand out from other entrants to the field of Cybersecurity.

Strategic Goal Support:

This goal covers my strategic goal of developing skills in AI development/management and advancing my technical knowledge with tools/skills used in the industry.

💼 3. Business Case & Justification

Professional Value:

This project will be added to a portfolio of projects I can showcase to employers.

Personal Value:

This project will help me develop a skill in AI development/management

Learning Outcomes:

I will learn technical skills related to AI, project management experience as a project manager, and deeper knowledge of the inner workings of AI model training.

📏 4. High-Level Scope

✅ In Scope for Semester Project

Determine which AI base model is most appropriate for this project
Secure training data to teach the model
Train the model off the training data and have it develop pattern recognition
Test the model with data not within the training dataset
Develop a presentation for end of class reporting to the professor

🗺️ Roadmap Items (≤6 Months)

Upload the model to allow usage beyond my personal server
Enhance the model with real data to refine its algorithms
Leverage the project for usage in my resume and job applications
Consider further applications and features that can be expanded on to create a more complex AI

❌ Out of Scope

Detecting subtle emotions such as "guilt" is more complex and most likely not accomplish-able in the duration of the semester.
Alternate Culture emotion detection can be difficult to achieve, as different groups of people can react differently when expressing emotion.
Long-term data collection is ethically questionable for the purposes of this educational project.
Integrating this model into wearable devices like smart glasses is a task too outside the requirements of just creating the AI.

Stakeholder	Category	Role/Interest	Power	Interest	Engagement Strategy
Dean of the college of Business and Economics	Internal	Non-invested party that is interested in the legal/moral alignment of the project	high	low	keep satisfied - ensure project does not violate academic standards of the college
Professor of Project Management class	Internal	Management overseer who will determine the value of the product	high	medium	Manage Closely - ensure project aligns with expectations and requirements set in the class
Sponsor/me	Internal	Most invested party seeking to create the project to meet class completion needs	high	high	manage closely - complete the project to specifications set in scope document
Legal Team	Internal	Ensure compliance to Legal regulations and proper usage	medium	low	keep informed - content review for compliance, policy adherence
AI model Developer	External	Involved party that created the base AI	medium	low	monitor - ensure project does not violate terms of use
Training data team	External	Involved party that are related to the AI model's training	low	low	monitor - ensure project does not violate privacy
Job Recruiters	External	Interested party for my technical acumen and skills	low	medium	monitor - align project with job recruitment key points
Clients	External	End user who wants to use the AI model	medium	high	keep informed - usability testing/training beyond initial dataset