We follow the same format as BANMo. We also save DINO features as lines of pixels to speed up dataloading.
DAVIS/
JPEGImages/
Full-Resolution/
sequence-name/
{%05d}.jpg
# segmentations from detectron2
Annotations/
Full-Resolution/
sequence-name/
{%05d}.png
# forward backward flow between every {1,2,4,8,16,32} frames from VCN-robust
FlowBW_%d/ and FlowFw_%d/
Full-Resolution/
sequence-name/ and optionally seqname-name_{%02d}/ (frame interval)
flo-{%05d}.pfm
occ-{%05d}.pfm
visflo-{%05d}.jpg
warp-{%05d}.jpg
# 16-dim Densepose features from CSE
Densepose/
Full-Resolution/
sequence-name/
# 112x(112*16) cropped densepose features
feat-{%05d}.pfm
# [x,y,w,h] saved to warp cropped features to original coordinate
bbox-{%05d}.txt
# densepose surface indices, for visualization
{%05d}.pfm
# lines of pixels in order to speed up dataloading
Pixels/
Full-Resolution/
sequence-name/
# skipped frames of flow followed by frame index
%d-%05d/
# lines of pixels for DINO features
DINO_Pixels/
Full-Resolution/
sequence-name/
# skipped frames of flow followed by frame index
%d-%05d/
# optical flow model
mkdir ./lasr_vcn
wget https://www.dropbox.com/s/bgsodsnnbxdoza3/vcn_rob.pth -O ./lasr_vcn/vcn_rob.pth
# PoseNet weights
mkdir -p mesh_material/posenet && cd "$_"
wget $(cat ../../misc/posenet.txt); cd ../../
Download the RigNet weights from here and unzip it under third_party/RigNet. Also create a folder to store the meshes as third_party/input_meshes
third_party/
RigNet/
checkpoints/
input_meshes/
other folders and files...
Download our iiwa dataset from here and place it at raw/iiwa.zip
# unzip the dataset
cd raw
unzip iiwa.zip
cd ..
# preprocess the dataset
bash preprocess/preprocess_iiwa.sh iiwa .mp4 24
# store as lines
seqname=iiwa
python preprocess/img2lines.py --seqname $seqname
# DINO features
python preprocess/prepare_dino_extract.py data_info_iiwa.pkl --seqname $seqname
python preprocess/compute_dino.py data_info_iiwa.pkl 16
Download swing and samba sequences from AMA dataset website or run the following scripts:
cd database; wget $(cat ../misc/ama.txt);
# untar files
ls *.tar | xargs -i tar xf {}
find ./T_* -type f -name "*.tgz" -execdir tar -xvzf {} \;
cd ../
Convert the data into correct format:
python scripts/ama-process/ama2davis.py --path ./database/T_samba
python scripts/ama-process/ama2davis.py --path ./database/T_swing
Then extract flow and dense appearance features:
seqname=ama-female
mkdir raw/$seqname;
# write filenames in replace of .MOV files
ls -d database/DAVIS/Annotations/Full-Resolution/T_s* | xargs -i echo {} | sed 's:.*/::' | xargs -i touch raw/$seqname/{}.txt # create empty txt files
bash preprocess/preprocess.sh $seqname .txt y 10
# store as lines
python preprocess/img2lines.py --seqname $seqname
To use DINO features, extract and save DINO features by running the following scripts:
python preprocess/prepare_dino_extract.py data_info_ama.pkl --seqname $seqname
python preprocess/compute_dino.py data_info_ama.pkl 16
First install soft rasterizer
pip install -e third_party/softras
Then download animated mesh sequences
mkdir database/eagle && cd "$_"
wget https://www.dropbox.com/sh/xz8kckfq817ggqd/AADIhtb1syWhDQeY8xa9Brc0a -O eagle.zip
unzip eagle.zip; cd ../../
Render image data and prepare mesh ground-truth
bash scripts/synthetic/render_eagle.sh
Store as lines and save DINO features
seqname=a-eagle
# store as lines
python preprocess/img2lines.py --seqname $seqname
# DINO features
python preprocess/prepare_dino_extract.py data_info_eagle.pkl --seqname $seqname
python preprocess/compute_dino.py data_info_eagle.pkl 16
Place all the .mp4 RGB videos under raw/my_data
# preprocess the dataset
bash preprocess/preprocess.sh my_data .mp4 n 24
# store as lines
seqname=my_data
python preprocess/img2lines.py --seqname $seqname
# DINO features
python preprocess/prepare_dino_extract.py data_info_my_data.pkl --seqname $seqname
python preprocess/compute_dino.py data_info_my_data.pkl 16