A skin-conformal rigid-in-soft array-based imaging system.

Area of Science:

• Robotics and Artificial Intelligence
• Materials Science and Engineering
• Computer Vision and Imaging

Background:

• Traditional array-based imaging systems lack flexibility, limiting their use with dynamic, non-rigid hosts like human skin or soft robots.
• Existing flexible visual sensor arrays have restricted pixel density and resolution due to fabrication constraints.

Purpose of the Study:

• To develop a skin-conformal imaging system, SkinSight, that achieves high-resolution imaging under complex mechanical deformations.
• To overcome the trade-off between compliance and resolution in flexible imaging systems.
• To enable robust, calibration-free 3D visual perception for augmented human and robotic capabilities.

Main Methods:

• Implemented a hybrid "rigid-in-soft" architecture integrating miniaturized imagers with stretchable interconnects.
• Utilized large-scale pre-trained vision transformers for end-to-end geometric inference to address processing challenges from continuous deformation.
• Developed a deformation-adaptive framework that does not rely on rigid structural priors.

Main Results:

• SkinSight achieves stable, high-resolution imaging on skin-conformal surfaces undergoing complex deformations.
• Demonstrated seamless integration with human skin, enabling a palm to act as a sensory interface.
• Successfully performed calibration-free 3D visual perception in dynamic environments.

Conclusions:

• SkinSight overcomes the limitations of rigid and flexible imaging systems, offering a high-resolution, compliant solution.
• The system augments perceptual capabilities for both humans and robots, showing potential for advanced wearables and embodied AI.
• This technology paves the way for next-generation human-computer interfaces and intelligent robotic systems.