Fig. 6: Hand-object reconstructions based on the ViTaM system. | Nature Communications

Fig. 6: Hand-object reconstructions based on the ViTaM system.

From: Capturing forceful interaction with deformable objects using a deep learning-powered stretchable tactile array

Fig. 6: Hand-object reconstructions based on the ViTaM system.

A The in-contact object reconstructions for two elastic sponges and a rigid scissor without and with strain interference suppression. B Three reconstruction stages of a gradually deformed dumpling-shaped plasticine manipulated by the hand without and with strain interference suppression. C A rigid folding rack reconstruction with multiple contacts by the hand on different spots on the object. D Visual-only and visual-tactile reconstruction results for a scissor, a rack, and a bottle, demonstrate the superiority of the visual-tactile joint learning. E Reconstruction sequence results of a deformable cup and a deformable sponge based on the visual-tactile data collected in the real world.

Back to article page