classroom-plus.html

Classroom+ (Nov 2018 - Jan 2019)

A real-time emotion recognition tool designed to help teachers
understand student engagement using just a standard webcam. It
captures facial snapshots and uses AI to estimate how students are
feeling -- no manual input, no surveys, minimal chance of error.

A teacher once thought her lessons were engaging, though many
students struggled to keep up and often felt disconnected. There
wasn't an easy way to give that kind of feedback without it feeling
personal or uncomfortable -- which led to building an objective
feedback tool for teachers.

== how it works ==

Set up a webcam at eye level. It automatically detects faces and
analyses expressions in real time, identifying one of six basic
emotions for each person.

== built with ==

Python 3.6.7, TensorFlow 1.15, OpenCV, NumPy, SciPy, Pandas, Haar
Cascades for face detection, and a custom-trained neural network
built with Keras for emotion classification.

Training data was self-created -- 600+ curated images from
volunteers at my school, manually sorted and labelled, and kept
private.

== looking back ==

This was a collaborative project, and I later found some of the
training logic had come from a GitHub repository I wasn't properly
credited from -- at the time I didn't understand what referencing
or attribution meant in open-source work. I'd handle credit and
licensing very differently if redoing it now.

Full write-up: https://tejastagra.com/projects/classroom-plus