Skip to content

ECG Arrhythmia Classification using Deep Learning

Classifies heartbeat types from ECG signals automatically.

Difficulty
Advanced
Build time
5 weeks
Tech stack
Python, WFDB, TensorFlow, SciPy
Team size
2–4 students
What's included
  • Complete source / working model
  • Project report in your college format
  • PPT for review & viva
  • Step-by-step setup guidance
  • Mentor support till your review
About this project
ECG records from the MIT-BIH database are filtered, segmented into individual beats and classified into normal and four arrhythmia classes using a 1D-CNN. The project compares accuracy with an SVM on hand-crafted features and visualises the beats the model gets wrong.
Modules
  1. 1 Signal processing Band-pass filtering and R-peak detection.
  2. 2 Classification 1D-CNN vs SVM comparison.
  3. 3 Evaluation Per-class metrics and error analysis.
Viva preview

Questions examiners ask on this project. Guided build students practise the full set with a mentor.

Q1What does each part of the PQRST wave represent?
Q2Why use a band-pass filter on ECG?
Q3Why is per-class recall important in medical AI?
Free

Get the abstract

We'll email it to you right away.

Book

Book this project

A mentor calls you within one working day to confirm scope, price and timeline. You pay after that call.