Retinomorphic Organic Infrared Imager Selectable between Frame-Based and Event-Driven Detection Modes

Area of Science:

• Optoelectronics
• Neuromorphic Engineering
• Image Sensors

Background:

• Traditional imaging systems often require separate components for motion detection and high-fidelity static image capture.
• Existing multimodal vision systems can be bulky and power-intensive due to complex circuitry.
• Retinomorphic sensors mimic biological vision but often lack reconfigurability for diverse tasks.

Purpose of the Study:

• To develop a single, reconfigurable retinomorphic infrared imager capable of operating in both motion and static modes.
• To demonstrate a device-level solution for switching between complementary imaging functionalities.
• To establish standardized measurement guidelines for time-derivative sensors.

Main Methods:

• A photocapacitor-transistor architecture with tunable variable resistance was employed for electrical mode switching.
• Device-level mode selection was implemented, eliminating the need for external circuit overhead.
• 8x8 pixel arrays were fabricated and integrated with silicon integrated circuits for spatiotemporal imaging.

Main Results:

• The imager successfully operated in two distinct modes: a motion mode with spiking outputs and a static mode with charge integration.
• Measurement guidelines for evaluating rate-dependent detection limits of time-derivative sensors were established.
• Fabricated arrays demonstrated scalability and integration capability for dual-mode spatiotemporal imaging.

Conclusions:

• The reconfigurable retinomorphic imager offers a compact and energy-efficient solution for multimodal vision.
• Device-level reconfigurability reduces circuit complexity and facilitates high-density integration.
• This technology paves the way for next-generation vision systems with adaptable sensing capabilities.