Matrix-CapsNet-EM-routing-tensorflow
This is a trial implementation of MATRIX CAPSULES WITH EM ROUTING in TensorFlow framework and Python programming language. （仅供交流学习使用 Education purpose only）
Project completeness 10%. To see a finsihed collaborative repository, see https://github.com/www0wwwjs1/Matrix-Capsules-EM-Tensorflow
Related Work
Link to CapsNet implementation （仅供交流学习使用 Education purpose only）: https://github.com/yhyu13/CapsNet-python-tensorflow
TODO
- Convert smallNORB dataset into TFRecords
- Split smallNORB dataset into manageable chunks (optional)
- Implement Convolutional Capsule Layers
- Implement EM routing
- Implement Spread Loss with linear annealing
- Reproduce test error on smallNORB dataset (if manageable)
- Reproduce experiments on smallNORB dataset (if manageable)
Dataset
This database is intended for experiments in 3D object reocgnition from shape. It contains images of 50 toys belonging to 5 generic categories: four-legged animals, human figures, airplanes, trucks, and cars. The objects were imaged by two cameras under 6 lighting conditions, 9 elevations (30 to 70 degrees every 5 degrees), and 18 azimuths (0 to 340 every 20 degrees).
The training set is composed of 5 instances of each category (instances 4, 6, 7, 8 and 9), and the test set of the remaining 5 instances (instances 0, 1, 2, 3, and 5).
Download NORB
cd data
./download.sh
The dataset will be available under folder smallNORB.
Write to TFRecord
python dataset.py tfrecord
The tfrecord files will appear under folder data.
Reference:
Dynamic Routing Between Capsules