A team from MIT’s Microsystems Technology Laboratory has created a chip that quickly and easily fixes common mistakes of smartphone images. In just a fraction of a second and using a lot less power, this processor can fix issues such as unsightly lighting and image noise that are common with smartphone cameras.
Funded by the Foxconn Technology Group based in Taiwan, the team designed a processor that is pretty impressive. And, if it works like they claim, you’ll be taking amazing smartphone images with just a tap of the screen.
This way-to-complex-for-most-of-us-to-truly-understand processor can handle the task of creating perfectly exposed images in a matter of milliseconds. How does this work? The processor takes 3 images: one underexposed, one correctly exposed and one overexposed and merges them together to form an image that captures details in both the dark and light areas of the picture. Also, the processor works with the phone’s camera flash to properly light an image. The chip takes two images: one with flash and one without. It merges these two images, preserving the ambient lighting from one image while using the details from the other. Finally, the processor eliminates noise from the image by using a bilateral filter to preserve outlines while selectively blurring pixels. The chip uses a complex bilateral grid in order for the filter to correctly blur pixels, yet not allowing it to blur across edges. All of these tasks will work together to create a more pleasing smartphone image.
The MIT team will be presenting their chip at the International Solid-State Circuits Conference in San Francisco this month. Good luck you guys! We hope you can rid the world of blown-out selfies, ugly food images and underexposed party pictures.