Ultrasensitive infrared-to-visible artificial vision via self-evolving projection guided by single-pixel detection
This paper presents SIVIS, a self-evolving projection system guided by single-pixel detection that achieves real-time, ultrasensitive infrared-to-visible upconversion with single-photon sensitivity, enabling direct naked-eye visualization of targets in extreme low-light conditions for applications ranging from night vision to biomedical imaging.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Human eyes are remarkably sensitive, yet they are blind to a vast portion of the light spectrum that surrounds us. We see the world in visible colors, but the infrared spectrum, which carries heat and chemical signatures, remains invisible to our natural vision. This limitation has long driven the development of cameras that can see in the dark or through smoke, but these devices often come with significant drawbacks. Traditional infrared cameras are expensive, bulky, and require complex cooling systems to function, making them difficult to shrink down for use in portable gear or wearable technology. Furthermore, many existing methods that try to convert invisible infrared light into visible light for the human eye struggle when light is scarce. They often demand intense laser power to work, which is impractical for delicate biological samples or situations where only a handful of photons are available.
Researchers have been seeking a way to bridge this gap, aiming to create a system that can take the faintest whispers of infrared light and turn them into a clear, visible image without the need for heavy equipment or blindingly bright lasers. The challenge lies in detecting these scarce photons and translating them into a picture that a person can see instantly. If such a system could be built, it would allow for night vision, medical imaging through tissue, and security checks that rely on seeing what is normally hidden, all while operating in conditions where light is almost non-existent.
A team of researchers has now developed a new approach called SIVIS, which stands for Self-evolving Infrared-to-VIsible upconversion with Single-pixel detection. This system represents a significant shift in how infrared vision is achieved. Instead of using a standard camera sensor with millions of tiny pixels to capture an image all at once, SIVIS uses a single, highly sensitive detector that acts like a single eye. This detector does not take a picture directly. Instead, it listens to the light bouncing off or passing through an object while a computer-controlled mirror rapidly flashes different patterns of light onto that object. The system is designed to be self-correcting. It starts with a random pattern of light, checks how much infrared light the single detector picks up, and then uses that information to adjust the next pattern. Over time, the patterns of light evolve, becoming more and more focused on the shape of the object, much like a sculptor chipping away stone until the form emerges.
The brilliance of this method is that it works in real-time and requires almost no light to function. The researchers demonstrated that their system can detect and visualize objects even when the infrared light is so faint that, on average, less than one photon hits each part of the object during a single moment of observation. To put this in perspective, the system operates at a power level roughly a hundred million times lower than what many other advanced infrared technologies require. As the light patterns evolve to match the object's shape, the system simultaneously projects a corresponding pattern of visible green light. This means that as the invisible infrared image is being reconstructed by the computer, a visible version of that same image is being painted onto a screen or the object itself, allowing a person to see the result with their naked eyes without any delay.
The team tested this technology in several different scenarios to prove its versatility. They showed that it could capture clear images of static objects, like a cutout shape, and track moving objects as they passed through the field of view. The system worked equally well in both transmission, where light passes through an object, and reflection, where light bounces off a surface. In one striking demonstration, the researchers used the system to reveal hidden security features printed with special ink that is invisible to the human eye under normal light. By projecting the visible patterns, the hidden text and shapes appeared instantly on the surface, effectively decrypting the invisible information.
Perhaps the most promising application shown in the study involves looking through obstacles. The researchers used the system to visualize vein-like structures hidden beneath a layer of artificial skin and even through a piece of raw chicken tissue. Because infrared light can penetrate biological tissue better than visible light, the system was able to detect the hidden structures and project a visible map of them onto a screen. This suggests that the technology could one day help doctors see blood vessels or other internal features without invasive procedures, simply by converting the infrared signals that pass through the body into a visible image.
The success of SIVIS relies on its ability to be incredibly efficient with the light it uses. By using a single-pixel detector that is sensitive enough to count individual photons, the system avoids the noise and inefficiency that plague traditional cameras in low-light conditions. The researchers found that even when the light was so dim that the detector was barely registering anything, the system could still slowly build up a clear picture of the target. This capability opens the door to a new generation of imaging tools that are compact, affordable, and capable of seeing in the deepest dark. While the current system takes a fraction of a second to evolve the image from a random pattern to a clear shape, the researchers note that future improvements in the algorithms could make this process even faster. For now, this work proves that it is possible to turn the faintest infrared signals into a bright, visible reality, offering a powerful new tool for seeing the unseen world.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.