NUS Enterprise

4-Eye Camera: A Method to Enhance Photographs by Transferring Texture and Contrast from Near Infrared (NIR) Images

Technology #09129n

Questions about this technology? Ask a Technology Manager

Download Printable PDF

Image Gallery
09129N 4-eye camera for enhanced quality of photographs_tech offer-NIR09129N 4-eye camera for enhanced quality of photographs_tech offer-RGB09129N 4-eye camera for enhanced quality of photographs_tech offer-Combined results
Mong Cheng Terence SIM
Managed By
Ms Asha Srinivasan (
Associate Director (65)65161671
Patent Protection
US Patent US 20100290703
Files and Attachments
Technology Offer [PDF]

In a typical outdoor scene, the ratio between the brightest and darkest regions of the scene --- called the dynamic range --- far exceeds what a camera can see.  This is true even for professional cameras. There is no single exposure that can capture all the details in the brightest and darkest regions simultaneously.  As a result, images omit a lot of scene details. By comparison, the human eye has a much higher dynamic range than a camera, and thus does not suffer from this problem as much.

The 4-eye camera uses Near-Infrared (NIR) light to augment the usual RGB colors, and is thus able to record more details.  In effect, this camera sees in 4 colors ("eyes"), hence its name.  By transferring the contrast and high-frequency information from NIR to RGB, this camera produces a single color photo that is visually more pleasing than the original color photo. Image details are enhanced while preserving the relative brightness/darkness of the scene. This software method could be built into the camera itself, or else provided as a software feature.

Figure 1.

Computer Vision

TRL 6. System/ subsystem model or prototype demonstration in a relevant environment.

To enhance quality of photos taken using amateur as well as professional cameras.

1. Only requires 2 images (near infrared + color photo) taken at the same time.
2. Works with moving objects.
3. Minimum hardware modifications to implement.
4. Fully automatic. No user intervention needed.
5. Adaptive. It handles different scenes, both outdoor and indoor, accordingly.
6. Fast. It takes only a few seconds to a minute to enhance images.
7. Non-machine learning: it does not require training data and is therefore no susceptible to the quality of training data.

Patent granted in the US. Available for licensing.