Overview
A deep-learning project that recognizes vehicles from photos.
The engineering challenge
Vehicle recognition requires learning discriminative visual features across many classes while dealing with variation in angles, lighting, and image quality.
What I built
- Developed a CNN-based image classification pipeline using TensorFlow and OpenCV.
- Worked with a dataset containing more than 60,000 labeled vehicle images.
- Used image-processing and data-preparation workflows to support model training and evaluation.
- Explored the performance and accuracy limitations of image-based classification.
What this demonstrates
Deep-learning fundamentals, image pipelines, dataset preparation, and model evaluation.