04 — Machine Learning
Pneumonia Detection.
- Category
- Machine Learning
- Stack
- Python, TensorFlow, Keras, MobileNetV2
Project 04Machine Learning
Pneumonia Detection
Python · TensorFlow · Keras · MobileNetV2
01
Overview
An image classification project exploring deep learning and transfer learning for chest X-ray image classification.
02
Project idea
Explore how deep learning and transfer learning can be applied to chest X-ray image classification.
03
Key features
- Transfer learning with MobileNetV2
- Chest X-ray image classification
- 3-block CNN experiment
04
Technologies
- Python
- TensorFlow
- Keras
- MobileNetV2
05
Architecture & workflow
- Step 01CNN experimentA 3-block CNN was built and trained to classify 32×32 RGB images across 10 categories.
- Step 02Transfer learningMobileNetV2 and transfer learning were applied to chest X-ray image classification.
06
Results & lessons learned
In the 10-category CNN experiment, the model reached 85% accuracy on 12,500 test samples. Chest X-ray classification metrics have not been published here yet.
The reported 85% accuracy (12,500 test samples) comes from a separate experiment: a 3-block CNN classifying 32×32 RGB images across 10 categories.
This is a learning project, not a clinically validated diagnostic tool.