Less verbose when extracting the base images