├── Dockerfile ├── README.md ├── datasets └── README.md └── notebook1.ipynb /Dockerfile: -------------------------------------------------------------------------------- 1 | # Start with a base docker image that has the modern deep learning libraries we need. 2 | FROM waleedka/modern-deep-learning 3 | MAINTAINER Waleed Abdulla 4 | 5 | # Install packages 6 | RUN apt-get install -y --no-install-recommends unzip 7 | 8 | # Clone the git repo into the /traffic directory. 9 | RUN git clone https://github.com/waleedka/traffic-signs-tensorflow /traffic 10 | 11 | # Download Belgian Traffic Dataset 12 | WORKDIR /traffic/datasets/ 13 | # Testing dataset 14 | RUN curl -o test.zip http://btsd.ethz.ch/shareddata/BelgiumTSC/BelgiumTSC_Testing.zip && \ 15 | unzip test.zip -d BelgiumTS/ && \ 16 | rm test.zip 17 | # Training dataset 18 | RUN curl -o train.zip http://btsd.ethz.ch/shareddata/BelgiumTSC/BelgiumTSC_Training.zip && \ 19 | unzip train.zip -d BelgiumTS/ && \ 20 | rm train.zip 21 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Traffic Signs Recognition with Tensorflow 2 | 3 | This is a multi-part tutorial to build a traffic signs recognition model using Tensorflow. 4 | 5 | This is a work in progress. Part 1 is included. Parts 2 and 3 are in development. 6 | 7 | -------------------------------------------------------------------------------- /datasets/README.md: -------------------------------------------------------------------------------- 1 | Download and expand the training and validation datasets here. Go to http://btsd.ethz.ch/shareddata/ and download the training and test data. There is a lot of datasets on that page, but you only need the two files listed under **BelgiumTS for Classification (cropped images)**": 2 | 3 | * BelgiumTSC_Training (171.3MBytes) 4 | * BelgiumTSC_Testing (76.5MBytes) 5 | 6 | After downloading and expanding the files, your directory structure should look something like this: 7 | 8 | ``` 9 | /traffic/datasets/BelgiumTS/Training/ 10 | /traffic/datasets/BelgiumTS/Testing/ 11 | ``` 12 | --------------------------------------------------------------------------------