Enhanced learning-based imaging with thin-film luminescent concentrators

Area of Science:

• Materials Science
• Computer Vision
• Electrical Engineering

Background:

• Traditional image sensors have limitations in flexibility and shape.
• Developing advanced human-computer interfaces requires innovative sensing technologies.

Purpose of the Study:

• To introduce LumiConSense, a transparent, flexible, and disposable thin-film image sensor.
• To present novel image reconstruction methods for enhanced sensor performance.
• To demonstrate real-time application potential through hardware and software prototypes.

Main Methods:

• Development of a new real-time image reconstruction algorithm.
• Integration of image reconstruction with shift-invariant linear image processing.
• Creation of hardware and software prototypes for LumiConSense.
• Design of a high-quality offline reconstruction algorithm.

Main Results:

• Significant enhancement in image quality using the new real-time reconstruction method.
• Efficient combination of image processing operations.
• Demonstration of real-time capabilities with functional prototypes.
• Achieved higher image quality with the offline reconstruction algorithm.

Conclusions:

• LumiConSense enables new human-computer interfaces with unprecedented shape and sensing flexibility.
• The developed reconstruction methods significantly improve image quality and processing efficiency.
• Prototypes confirm the sensor's viability for real-time applications.