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