Event
Wearable Robot Grasp-Part Segmentation Extended to Multi-Object Scenes
1 reports · 1 sources
Simone Lugani et al. propose adding a laser pointer to a vision-based graspable-part recognition pipeline for target disambiguation in multi-object scenes, and use a mixed-reality strategy to reuse existing high-quality segmentation masks to generate training images and retrain a deep convolutional neural network, in order to support grasp selection and pre-shaping for upper-limb prostheses. Experimental results show that the method improves target disambiguation accuracy by up to 9% compared with existing approaches while keeping computational cost unchanged.
Report timeline
- Extending graspable-part segmentation to multi-object images
A study proposes introducing a laser pointer into a vision-based portable pipeline for target disambiguation in multi-object scenes, to support grasp selection and preshaping in upper-limb prostheses. The method uses a mixed reality strategy to reuse existing high-quality segmentation masks to generate training images and retrains a deep convolutional neural network. Experimental results show target disambiguation accuracy improves by up to 9% over existing solutions while computational cost remains unchanged.
arXiv cs.RO ·