Event-based Grasping Datsets

The datasets for the works 'Event-based Robotic Grasping Detection with Neuromorphic Vision Sensor and Event-Grasping Dataset'[Link] and 'NeuroGrasp: Multimodal Neural Network With Euler Region Regression for Neuromorphic Vision-Based Grasp Pose Estimation'[Link]

Grasping DataSets Captured by DAVIS (Dynamic vision sensor)

1. Event-based Robotic Grasping Detection with Neuromorphic Vision Sensor and Event-Grasping Dataset

This is the implementation of our recently work 'Event-Based Robotic Grasping Detection With Neuromorphic Vision Sensor and Event-Grasping Dataset'. We construct a robotic grasping dataset name E-Grapsing with 91 objects. The dataset can be donwload from this link(https://drive.google.com/file/d/1frAvxFjjhBzhhNd2SeDKN1R6E417jD3L/view?usp=sharing)

If you find it helpful for your research, please cite:

@ARTICLE{10.3389/fnbot.2020.00051,
  
AUTHOR={Li, Bin and Cao, Hu and Qu, Zhongnan and Hu, Yingbai and Wang, Zhenke and Liang, Zichen},   
	 
TITLE={Event-Based Robotic Grasping Detection With Neuromorphic Vision Sensor and Event-Grasping Dataset},      
	
JOURNAL={Frontiers in Neurorobotics},      
	
VOLUME={14},      

PAGES={51},     
	
YEAR={2020},      
	  
URL={https://www.frontiersin.org/article/10.3389/fnbot.2020.00051},       
	
DOI={10.3389/fnbot.2020.00051},      
	
ISSN={1662-5218},   
}

1.1 Single Grasp DataSet

We represent the tracked LED points with rectangles, each event frame acquires the ground truth of a good grasping position, as is shown in below

1.2 Multi-Grasp DataSet

We manually set ground truth annotations in the first frame, the transformed annotations in any other frame canbe attained. The final dataset with multi-grasping annotations is illustrated in below

2. NeuroGrasp: Multimodal Neural Network With Euler Region Regression for Neuromorphic Vision-Based Grasp Pose Estimation

This is the implementation of our recently work 'NeuroGrasp: Multimodal Neural Network With Euler Region Regression for Neuromorphic Vision-Based Grasp Pose Estimation'. We construct a neuromorphic vision-based robotic grasp dataset with 154 moving objects, named NeuroGrasp, which is the first RGB-Event multimodality grasp dataset (to the best of our knowledge). The dataset can be donwload from this link(https://drive.google.com/file/d/1YVBJct2aZmugWolKTI-d-zUoVKnViIb6/view?usp=sharing)

If you find it helpful for your research, please cite:

@ARTICLE{9787342,
  author={Cao, Hu and Chen, Guang and Li, Zhijun and Hu, Yingbai and Knoll, Alois},
  journal={IEEE Transactions on Instrumentation and Measurement}, 
  title={NeuroGrasp: Multimodal Neural Network With Euler Region Regression for Neuromorphic Vision-Based Grasp Pose Estimation}, 
  year={2022},
  volume={71},
  number={},
  pages={1-11},
  doi={10.1109/TIM.2022.3179469}}