← Back to Projects
CNN Image Classification Lab
PythonPyTorchtorchvisionNumPyMatplotlibCNNsTransfer LearningData Augmentation
An image classification system developed as part of the EDAP30 Advanced Applied Machine Learning course. The project implements both custom CNN architectures and transfer learning approaches using. The system achieves high accuracy in classifying flower species from the Flowers Dataset.
Key Features
- ✓Implementation of custom CNN architecture
- ✓Transfer learning implementation using pre-trained ResNet50
- ✓Comprehensive data preprocessing and augmentation pipeline
Challenges & Solutions
- ↳Finding good augmentation techniques to improve model generalization
- ↳Designing an efficient custom CNN architecture from scratch
- ↳Implementing and fine-tuning transfer learning with ResNet50
Project Screenshots