Skip to content
/ HaDR Public
forked from WXinlong/SOLO

Hand instance segmentation using synthetic dataset

License

Notifications You must be signed in to change notification settings

anion0278/HaDR

 
 

Repository files navigation

HaDR: Applying Domain Randomization for Generating Synthetic Multimodal Dataset for Hand Instance Segmentation in Cluttered Industrial Environments

Installation

Requirements

  • Linux (Windows is not officially supported)
  • Python 3.5+
  • PyTorch 1.1 or higher (>=1.5 is not tested)
  • CUDA 9.0 or higher
  • NCCL 2
  • GCC 4.9 or higher
  • mmcv 0.2.16

Install

a. Create a conda virtual environment and activate it.

conda create -n mmdet python=3.7 -y
conda activate mmdet

b. Install PyTorch and torchvision following the official instructions, e.g.,

conda install pytorch torchvision -c pytorch

c. Clone the repository.

git clone https://github.com/anion0278/HaDR.git
cd HaDR

d. Install build requirements and then install.

pip install -r requirements/build.txt
pip install "git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI"
pip install -v -e .  # or "python setup.py develop"

Dataset and pretrained models

The dataset and pretrained SOLOv2 and Mask R-CNN models can be found on Kaggle.

Demo

demo

Languages

  • Python 72.3%
  • C++ 16.7%
  • Cuda 10.9%
  • Cython 0.1%
  • Shell 0.0%
  • C 0.0%