RoryShao / cnn_car_classification

car classification based on cnn

CNN based Car Classification

CNN is widely used in image classification. In this project, we are going to classify car images.

dataset

stanford car dataset from https://ai.stanford.edu/~jkrause/cars/car_dataset.html

How to use

download car dataset, prepare the data for the model.

1. download training image http://imagenet.stanford.edu/internal/car196/cars_train.tgz uncompress to ./data/cars_train
2. download testing image http://imagenet.stanford.edu/internal/car196/cars_test.tgz uncompress to ./data/cars_test
3. download devkit https://ai.stanford.edu/~jkrause/cars/car_devkit.tgz uncompress to ./data/devkit
4. download test annotation with class label http://imagenet.stanford.edu/internal/car196/cars_test_annos_withlabels.mat move it to ./data/devkit
5. (optional) download bounding box annotations for all images http://imagenet.stanford.edu/internal/car196/cars_annos.mat move it to ./data/devkit
6. run data_prepare.py to prepare the training and testing data for the model (you may need modify some paths in the file)

Training

run python train.py -t /path/to/car_dataset/train/ -v /path/to/car_dataset/test/ -m vgg16 -s car196 -e 20 -n 196