GitPlanet

Cheap and reliable Node.js hosting starts at $3/month, and $1/month static HTML hosting

Created with love in Canada, visit hostnodejs.com today

Feel like to post an Ad? Learn Details
All Projects → anibali → H36m Fetch

anibali / H36m Fetch

Licence: apache-2.0
Human 3.6M 3D human pose dataset fetcher

Programming Languages

python
139335 projects - #7 most used programming language

Labels

datasethuman-pose-estimation

Projects that are alternatives of or similar to H36m Fetch

Evoskeleton
Official project website for the CVPR 2020 paper (Oral Presentation) "Cascaded Deep Monocular 3D Human Pose Estimation With Evolutionary Training Data"
Stars: ✭ 154 (-30%)
Mutual labels:  dataset, human-pose-estimation
Human3.6m downloader
Human3.6M downloader by Python
Stars: ✭ 37 (-83.18%)
Mutual labels:  dataset, human-pose-estimation
Split Folders
🗂 Split folders with files (i.e. images) into training, validation and test (dataset) folders
Stars: ✭ 203 (-7.73%)
Mutual labels:  dataset
Datatable
A go in-memory table
Stars: ✭ 215 (-2.27%)
Mutual labels:  dataset
Omnianomaly
KDD 2019: Robust Anomaly Detection for Multivariate Time Series through Stochastic Recurrent Neural Network
Stars: ✭ 208 (-5.45%)
Mutual labels:  dataset
Covid19za
Coronavirus COVID-19 (2019-nCoV) Data Repository and Dashboard for South Africa
Stars: ✭ 208 (-5.45%)
Mutual labels:  dataset
Ava downloader
⏬ Download AVA dataset (A Large-Scale Database for Aesthetic Visual Analysis)
Stars: ✭ 214 (-2.73%)
Mutual labels:  dataset
Semantic Segmentation Suite
Semantic Segmentation Suite in TensorFlow. Implement, train, and test new Semantic Segmentation models easily!
Stars: ✭ 2,395 (+988.64%)
Mutual labels:  dataset
Cu Net
Code for "Quantized Densely Connected U-Nets for Efficient Landmark Localization" (ECCV 2018) and "CU-Net: Coupled U-Nets" (BMVC 2018 oral)
Stars: ✭ 218 (-0.91%)
Mutual labels:  human-pose-estimation
Dynamic Training Bench
Simplify the training and tuning of Tensorflow models
Stars: ✭ 210 (-4.55%)
Mutual labels:  dataset
Dataset Serialize
JSON to DataSet and DataSet to JSON converter for Delphi and Lazarus (FPC)
Stars: ✭ 213 (-3.18%)
Mutual labels:  dataset
Charlatan
Create fake data in R
Stars: ✭ 209 (-5%)
Mutual labels:  dataset
Computervisiondatasets
Stars: ✭ 207 (-5.91%)
Mutual labels:  dataset
Short Jokes Dataset
Python scripts for building 'Short Jokes' dataset, featured on Kaggle
Stars: ✭ 215 (-2.27%)
Mutual labels:  dataset
Pytorch realtime multi Person pose estimation
Pytorch version of Realtime Multi-Person Pose Estimation project
Stars: ✭ 205 (-6.82%)
Mutual labels:  human-pose-estimation
Bccd dataset
BCCD (Blood Cell Count and Detection) Dataset is a small-scale dataset for blood cells detection.
Stars: ✭ 216 (-1.82%)
Mutual labels:  dataset
Tech.ml.dataset
A Clojure high performance data processing system
Stars: ✭ 205 (-6.82%)
Mutual labels:  dataset
Binary Human Pose Estimation
This code implements a demo of the Binarized Convolutional Landmark Localizers for Human Pose Estimation and Face Alignment with Limited Resources paper by Adrian Bulat and Georgios Tzimiropoulos.
Stars: ✭ 210 (-4.55%)
Mutual labels:  human-pose-estimation
Pynasa
Stars: ✭ 212 (-3.64%)
Mutual labels:  dataset
Pyranet
Code for "Learning Feature Pyramids for Human Pose Estimation" (ICCV 2017)
Stars: ✭ 222 (+0.91%)
Mutual labels:  human-pose-estimation
View All Similar Projects ➔

Human3.6M dataset fetcher

Human3.6M is a 3D human pose dataset containing 3.6 million human poses and corresponding images. The scripts in this repository make it easy to download, extract, and preprocess the images and annotations from Human3.6M.

Please do not ask me for a copy of the Human3.6M dataset. I do not own the data, nor do I have permission to redistribute it. Please visit http://vision.imar.ro/human3.6m/ in order to request access and contact the maintainers of the dataset.

Requirements

  • Python 3
  • axel
  • CDF
  • ffmpeg 3.2.4

Alternatively, a Dockerfile is provided which has all of the requirements set up. You can use it to run scripts like so:

$ docker-compose run --rm --user="$(id -u):$(id -g)" main python3 <script>

Usage

  1. Firstly, you will need to create an account at http://vision.imar.ro/human3.6m/ to gain access to the dataset.
  2. Once your account has been approved, log in and inspect your cookies to find your PHPSESSID.
  3. Copy the configuration file config.ini.example to config.ini and fill in your PHPSESSID.
  4. Use the download_all.py script to download the dataset, extract_all.py to extract the downloaded archives, and process_all.py to preprocess the dataset into an easier to use format.

Frame sampling

Not all frames are selected during the preprocessing step. We assume that the data will be used in the Protocol #2 setup (see "Compositional Human Pose Regression"), so for subjects S9 and S11 every 64th frame is used. For the training subjects (S1, S5, S6, S7, and S8), only "interesting" frames are used. That is, near-duplicate frames during periods of low movement are skipped.

You can edit select_frame_indices_to_include() in process_all.py to change this behaviour.

License

The code in this repository is licensed under the terms of the Apache License, Version 2.0.

Please read the license agreement for the Human3.6M dataset itself, which specifies citations you must make when using the data in your own research. The file metadata.xml is directly copied from the "Visualisation and large scale prediction software" bundle from the Human3.6M website, and is subject to the same license agreement.

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].