The Anki Cozmo robot can recognize faces and objects with markers like Cozmo’s Power Cubes. Ansgar Schmidt and I have developed some sample code to allow Cozmo to recognize other types of objects via TensorFlow.
Watch the video and check out the slides to see how Cozmo can recognize three different toys:
Get the source code from GitHub.
The training is done via TensorFlow and a retrained MobileNet model on Kubernetes.
The classification is done via Tensorflow running in an OpenWhisk function.
Ansgar and I have blogged about this project in more detail.
- Visual Recognition with TensorFlow and OpenWhisk
- Accessing IBM Object Storage from Python
- Image Recognition with Tensorflow training on Kubernetes
- Image Recognition with Tensorflow classification on OpenWhisk
- Sample Application to classify Images with TensorFlow and OpenWhisk
Cozmo finding the deer:
If you want to run this demo, you can get an account on the IBM Cloud and try OpenWhisk and Kubernetes.