Skip to content
Tensorflow tutorial in ROS
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
model
LICENSE.txt
README.md Update README.md May 19, 2017
input_data.py Tensorflow version update May 12, 2017
tensorflow_in_ros_mnist.py changing code for @shixinlishixinli issue May 19, 2017
train.py Tensorflow version update May 12, 2017

README.md

Tensorflow_in_ROS

I made a ros-node to predict numbers from camera image.
I used Tensorflow tutorial Deep MNIST model(https://www.tensorflow.org/versions/r0.11/tutorials/mnist/pros/index.html)

Before you use this, you must install some programs.

$ sudo apt-get install ros-indigo-cv-bridge ros-indigo-cv-camera

tensorflow_in_ros_mnist.py

This is main file. This make the ros-node.

  • publish: /result (std_msgs/Int16)
  • subscribe: /image (sensor_msgs/Image)

model.ckpt,model.ckpt.meta

This is Trained model. Test set accuracy is approximately 99.2%.

How to try

$ roscore
$ python tensorflow_in_ros_mnist.py image:=/cv_camera/image_raw
$ rosrun cv_camera cv_camera_node
$ rostopic echo /result

Result

You can see ros-node published predicted numbers. Look at this.
Right picture is 9-image from camera.
Left number is published predicted number.

Result

Reference

This software is released under the Apache License2.0, see LICENSE.txt.

You can’t perform that action at this time.