Martin Sundermeyer, Zoltan-Csaba Marton, Maximilian Durner, Manuel Brucker, Rudolph Triebel
Best Paper Award, ECCV 2018.

paper, supplement, oral

If you find Augmented Autoencoders useful for your research, please consider citing:

@InProceedings{Sundermeyer_2018_ECCV,
author = {Sundermeyer, Martin and Marton, Zoltan-Csaba and Durner, Maximilian and Brucker, Manuel and Triebel, Rudolph},
title = {Implicit 3D Orientation Learning for 6D Object Detection from RGB Images},
booktitle = {The European Conference on Computer Vision (ECCV)},
month = {September},
year = {2018}
}

Martin Sundermeyer, Maximilian Durner, En Yen Puang, Zoltan-Csaba Marton, Narunas Vaskevicius, Kai O. Arras, Rudolph Triebel
CVPR 2020
The code of this work can be found here

We propose a real-time RGB-based pipeline for object detection and 6D pose estimation. Our novel 3D orientation estimation is based on a variant of the Denoising Autoencoder that is trained on simulated views of a 3D model using Domain Randomization. This so-called Augmented Autoencoder has several advantages over existing methods: It does not require real, pose-annotated training data, generalizes to various test sensors and inherently handles object and view symmetries.

1. Train the Augmented Autoencoder(s) using only a 3D model to predict 3D Object Orientations from RGB image crops
2. For full RGB-based 6D pose estimation, also train a 2D Object Detector (e.g. https://github.com/fizyr/keras-retinanet)
3. Optionally, use our standard depth-based ICP to refine the 6D Pose

Nvidia GPU with >4GB memory (or adjust the batch size)
RAM >8GB
Duration depending on Configuration and Hardware: ~3h per Object

Python 3.6+
CUDA 11.0: https://developer.nvidia.com/cuda-11.0-download-archive
cuDNN: https://developer.nvidia.com/rdp/cudnn-download
VOC training set (mirror): https://pjreddie.com/media/files/VOCtrainval_11-May-2012.tar

sudo apt-get install libglfw3-dev libglfw3  

sudo apt-get install libassimp-dev  

Continued in the the All Platforms section below.

1. Download from https://www.glfw.org/
   Tested with Version 3.3.3 (https://github.com/glfw/glfw/releases/download/3.3.3/glfw-3.3.3.bin.WIN64.zip)
2. Unzip to a permanent location
3. Set the GLFW_ROOT environment variable to the location where it was unzipped
4. Install cython & wheel

   pip install --user cython wheel

5. Install the pip package

   pip install --user cyglfw3

   If you get an error about missing Visual Studio build tools, you may need to install them through the Visual Studio installer.
6. Copy the file glfw3.dll from lib-vc2012 into the Python Lib\site-packages\cyglfw3 where cyglfw is installed.
7. Add the lib-vc2012 directory to the PATH environment variable

1. Download & Install the package from https://github.com/assimp/assimp/releases/tag/v4.1.0/
2. Add the <Program Files>\Assimp\bin\x64 directory to the PATH environment variable

If you run into the error "Unable to load numpy_formathandler accelerator from OpenGL_accelerate", replace numpy with numpy‑1.19.5+mkl‑cp38‑cp38‑win_amd64.whl from https://www.lfd.uci.edu/~gohlke/pythonlibs/#numpy (run "pip install" on the whl). Match the value for cp38 to your Python version - i.e. cp38 is for Python 3.8.

Continued in the the All Platforms section below.

1. See https://pypi.org/project/opencv-python/ for GPU acceleration instructions.
   To install the CPU-only version:

pip install --user opencv-python

2a) Latest versions of dependencies:

pip install --user --pre --upgrade PyOpenGL PyOpenGL_accelerate
pip install --user tensorflow cython cyglfw3 pyassimp==3.3 imgaug progressbar tf_agents

2b) Exact versions of dependencies:

pip install --user tensorflow==2.4.1 PyOpenGL==3.1.5 PyOpenGL_accelerate==3.1.5 cython==0.29.19 cyglfw3==3.1.0.2 pyassimp==3.3 imgaug==0.4.0 progressbar==2.5 tf_agents==0.7.1

Please note that we use the GLFW context as default which does not support headless rendering. To allow for both, onscreen rendering & headless rendering on a remote server, set the context to EGL:

export PYOPENGL_PLATFORM='egl'

In order to make the EGL context work, you might need to change PyOpenGL like here

1. Pip installation

pip install --user .

2. Set Workspace path

Consider adding this to your bash profile

export AE_WORKSPACE_PATH=/path/to/autoencoder_ws  

Edit your system environment variables to add the AE_WORKSPACE_PATH variable which points to the location where you want to store training results and configurations for AugmentedAutoencoder

3. Create Workspace, Init Workspace (if installed locally, make sure .local/bin/ is in your PATH)

mkdir $AE_WORKSPACE_PATH
cd $AE_WORKSPACE_PATH
ae_init_workspace

1. Create the training config file. Insert the paths to your 3D model and background images.

mkdir $AE_WORKSPACE_PATH/cfg/exp_group
cp $AE_WORKSPACE_PATH/cfg/train_template.cfg $AE_WORKSPACE_PATH/cfg/exp_group/my_autoencoder.cfg
gedit $AE_WORKSPACE_PATH/cfg/exp_group/my_autoencoder.cfg

2. Generate and check training data. The object views should be strongly augmented but identifiable.

(Press ESC to close the window.)

ae_train exp_group/my_autoencoder -d

This command does not start training and should be run on a PC with a display connected.

Output:

3. Train the model (See the Headless Rendering section if you want to train directly on a server without display)

ae_train exp_group/my_autoencoder

$AE_WORKSPACE_PATH/experiments/exp_group/my_autoencoder/train_figures/training_images_29999.png  

Middle part should show reconstructions of the input object (if all black, set higher bootstrap_ratio / auxilliary_mask in training config)

4. Create the embedding

ae_embed exp_group/my_autoencoder

have a look at /auto_pose/test/

Feed one or more object crops from disk into AAE and predict 3D Orientation

python aae_image.py exp_group/my_autoencoder -f /path/to/image/file/or/folder

The same with a webcam input stream

python aae_webcam.py exp_group/my_autoencoder

Option 1: Train a RetinaNet Model from https://github.com/fizyr/keras-retinanet

adapt $AE_WORKSPACE_PATH/eval_cfg/aae_retina_webcam.cfg

python auto_pose/test/aae_retina_webcam_pose.py -test_config aae_retina_webcam.cfg -vis

Option 2: Using the Google Detection API with Fixes

Train a 2D detector following https://github.com/naisy/train_ssd_mobilenet
adapt /auto_pose/test/googledet_utils/googledet_config.yml

python auto_pose/test/aae_googledet_webcam_multi.py exp_group/my_autoencoder exp_group/my_autoencoder2 exp_group/my_autoencoder3

For the evaluation you will also need https://github.com/thodan/sixd_toolkit + our extensions, see sixd_toolkit_extension/help.txt

Create the evaluation config file

mkdir $AE_WORKSPACE_PATH/cfg_eval/eval_group
cp $AE_WORKSPACE_PATH/cfg_eval/eval_template.cfg $AE_WORKSPACE_PATH/cfg_eval/eval_group/eval_my_autoencoder.cfg
gedit $AE_WORKSPACE_PATH/cfg_eval/eval_group/eval_my_autoencoder.cfg

Set estimate_bbs=False in the evaluation config

ae_eval exp_group/my_autoencoder name_of_evaluation --eval_cfg eval_group/eval_my_autoencoder.cfg
e.g.
ae_eval tless_nobn/obj5 eval_name --eval_cfg tless/5.cfg

Set estimate_bbs=True in the evaluation config

Generate a training dataset for T-Less using detection_utils/generate_sixd_train.py

python detection_utils/generate_sixd_train.py

Train https://github.com/fizyr/keras-retinanet or https://github.com/balancap/SSD-Tensorflow

ae_eval exp_group/my_autoencoder name_of_evaluation --eval_cfg eval_group/eval_my_autoencoder.cfg
e.g.
ae_eval tless_nobn/obj5 eval_name --eval_cfg tless/5.cfg

[Paths]
# Path to the model file. All formats supported by assimp should work. Tested with ply files.
MODEL_PATH: /path/to/my_3d_model.ply
# Path to some background image folder. Should contain a * as a placeholder for the image name.
BACKGROUND_IMAGES_GLOB: /path/to/VOCdevkit/VOC2012/JPEGImages/*.jpg

[Dataset]
#cad or reconst (with texture)
MODEL: reconst
# Height of the AE input layer
H: 128
# Width of the AE input layer
W: 128
# Channels of the AE input layer (default BGR)
C: 3
# Distance from Camera to the object in mm for synthetic training images
RADIUS: 700
# Dimensions of the renderered image, it will be cropped and rescaled to H, W later.
RENDER_DIMS: (720, 540)
# Camera matrix used for rendering and optionally for estimating depth from RGB
K: [1075.65, 0, 720/2, 0, 1073.90, 540/2, 0, 0, 1]
# Vertex scale. Vertices need to be scaled to mm
VERTEX_SCALE: 1
# Antialiasing factor used for rendering
ANTIALIASING: 8
# Padding rendered object images and potentially bounding box detections 
PAD_FACTOR: 1.2
# Near plane
CLIP_NEAR: 10
# Far plane
CLIP_FAR: 10000
# Number of training images rendered uniformly at random from SO(3)
NOOF_TRAINING_IMGS: 10000
# Number of background images that simulate clutter
NOOF_BG_IMGS: 10000

[Augmentation]
# Using real object masks for occlusion (not really necessary)
REALISTIC_OCCLUSION: False
# Maximum relative translational offset of input views, sampled uniformly  
MAX_REL_OFFSET: 0.20
# Random augmentations at random strengths from imgaug library
CODE: Sequential([
    #Sometimes(0.5, PerspectiveTransform(0.05)),
    #Sometimes(0.5, CropAndPad(percent=(-0.05, 0.1))),
    Sometimes(0.5, Affine(scale=(1.0, 1.2))),
    Sometimes(0.5, CoarseDropout( p=0.2, size_percent=0.05) ),
    Sometimes(0.5, GaussianBlur(1.2*np.random.rand())),
    Sometimes(0.5, Add((-25, 25), per_channel=0.3)),
    Sometimes(0.3, Invert(0.2, per_channel=True)),
    Sometimes(0.5, Multiply((0.6, 1.4), per_channel=0.5)),
    Sometimes(0.5, Multiply((0.6, 1.4))),
    Sometimes(0.5, ContrastNormalization((0.5, 2.2), per_channel=0.3))
    ], random_order=False)

[Embedding]
# for every rotation save rendered bounding box diagonal for projective distance estimation
EMBED_BB: True
# minimum number of equidistant views rendered from a view-sphere
MIN_N_VIEWS: 2562
# for each view generate a number of in-plane rotations to cover full SO(3)
NUM_CYCLO: 36

[Network]
# additionally reconstruct segmentation mask, helps when AAE decodes pure blackness
AUXILIARY_MASK: False
# Variational Autoencoder, factor in front of KL-Divergence loss
VARIATIONAL: 0
# Reconstruction error metric
LOSS: L2
# Only evaluate 1/BOOTSTRAP_RATIO of the pixels with highest errors, produces sharper edges
BOOTSTRAP_RATIO: 4
# regularize norm of latent variables
NORM_REGULARIZE: 0
# size of the latent space
LATENT_SPACE_SIZE: 128
# number of filters in every Conv layer (decoder mirrored)
NUM_FILTER: [128, 256, 512, 512]
# stride for encoder layers, nearest neighbor upsampling for decoder layers
STRIDES: [2, 2, 2, 2]
# filter size encoder
KERNEL_SIZE_ENCODER: 5
# filter size decoder
KERNEL_SIZE_DECODER: 5


[Training]
OPTIMIZER: Adam
NUM_ITER: 30000
BATCH_SIZE: 64
LEARNING_RATE: 1e-4
SAVE_INTERVAL: 5000

[Queue]
# number of threads for producing augmented training data (online)
NUM_THREADS: 10
# preprocessing queue size in number of batches
QUEUE_SIZE: 50