Tokyo scientists create self-powered artificial synapse for color-based machine vision

New device from Tokyo University of Science replicates human-like color recognition using solar-powered neuromorphic computing.

Image credit: Associate Professor Takashi Ikuno from Tokyo University of Science, Testing a new artificial synapse for computer vision tasks

Researchers at Tokyo University of Science (TUS) have developed an artificial synapse capable of performing color-based image recognition without the need for external power. The device integrates dye-sensitized solar cells to generate electricity and can distinguish colors across the visible spectrum with a resolution of 10 nanometers.

This development addresses two longstanding barriers in computer vision: achieving high-precision color detection and reducing energy demands. The technology mimics the human visual system, offering an alternative to current machine vision setups that consume significant computing and power resources—especially at the edge, where devices like smartphones or drones must operate independently.

TUS is Japan’s largest private research university focused on science. It conducts multidisciplinary research in areas such as nanomaterials, electronics, and artificial intelligence.

Technical details and functionality

The new device combines two types of dye-sensitized solar cells that respond to different wavelengths of light, producing opposite voltage polarities under red and blue illumination. This allows the system to execute logic operations without additional circuitry.

In testing, the synapse showed the ability to detect subtle differences in color and generate both positive and negative voltages in response to specific wavelengths. According to the researchers, this dual response can replace more complex hardware currently used in visual processing systems.

Takashi Ikuno, associate professor at TUS and lead author of the study, says, “The results show great potential for the application of this next-generation optoelectronic device, which enables high-resolution color discrimination and logical operations simultaneously, to low-power artificial intelligence (AI) systems with visual recognition.”

Real-world testing and performance

To validate its capabilities, the team used the device within a reservoir computing framework to classify color-coded human motion data. The system was able to recognize 18 different color-motion combinations with an accuracy rate of 82 percent, using only a single synapse-based device, as opposed to the multiple photodiodes required in traditional approaches.

Use cases extend across several industries. In automotive systems, the technology could support color-based recognition of signals and signage with minimal battery drain. In healthcare, it may enable wearables that monitor patient vitals like oxygen saturation without frequent recharging. In consumer electronics, battery-efficient smartphones or AR headsets could benefit from embedded machine vision without heavy processing overhead.

Ikuno adds, “We believe this technology will contribute to the realization of low-power machine vision systems with color discrimination capabilities close to those of the human eye, with applications in optical sensors for self-driving cars, low-power biometric sensors for medical use, and portable recognition devices.”