A Computational Approach for Obstruction-Free Photography

The video accompanying our SIGGRAPH 2015 paper " A Computational Approach for Obstruction-Free Photography". We present a unified computational approach for taking photos through reflecting or occluding elements such as windows and fences. Rather than capturing a single image, we instruct the user to take a short image sequence while slightly moving the camera. Differences that often exist in the relative position of the background and the obstructing elements from the camera allow us to separate them based on their motions, and to recover the desired background scene as if the visual obstructions were not there. We show results on controlled experiments and many real and practical scenarios, including shooting through reflections, fences, and raindrop-covered windows.

[MikeDownes]

pretty excellent example, what i like the most - recovery of the reflected image .. that has some brilliant problem solving applications ..

[jjvictor]

This is one of the most impressive things i've ever seen in my life, congratulations for all the team behind this.

[alexsantana]

Finally, been waiting for some genius to make something like this.

[BrainSeepsOut]

The recovered reflections make me think about CSI! "ENHANCE!"

[xZerplinxProduction]

this literally makes those ridiculous csi moments possible, good job

[readneuromancerbywilliamgi]

This tech is going to be worth a lot of money. I would not sell this algorithm to anyone ever. This can make killing in licensing. Photo apps will buy this service from you as a feature for their video / photo correction tools. Production Houses will use this tech to fix up shots made in rough weather conditions. Government agencies may even use this tech to enhance surveillance footage. Adobe might purchase usage for AfterEffects / Photoshop. 

Just a thought. Don't sell your algorithm, it's too incredibly versatile and opens up a world of business options that an entire fortune can be built on. Great work.

[habeebomotunde]

it's amazing what you can achieve with this kind of functionality. Just a day before, this use case would have solved a problem i encountered while taking a photo through a dense car window. The reflection really messed up the main image but i'm sure this really comes in handy. good job guys

[gadgetsgimmicksandtech]

This is just amazing.  Never mind enhancement for traditional photographs, this has to be of use in forensic detection!

[iLikeTheUDK]

This should be made available for some visual effects vendors. It's going to allow some amazing new things.

[AlexHannemanVFX]

You guys should release some open-source software for this!

[carlospozo]

Amazing, this teach us that in a mear figure we will use bionic eyes yo solve problema like reflection ... A very good work, congratulations

[TheBitProgress]

Wow, that's cool. Motion flow + 3D Separation. I think as side effect you can get noise reduction and detail's reconstruction.

[gavinator]

Incredible work everyone. Really amazing. You should all be very proud. :)

[nerwin]

This is really impressive. I wonder if something like this already exist in terms of digital forensics maybe not so much removing the reflection, but enhancing the visibility of the reflection for identification purposes. Either way, this is incredible. Nice work.

[sfernangarroj]

Gran avance, ojalá lo apliquen pronto

[MyLittleMagneton]

This could be a game changer for gaining evidence at crime scenes.

[TulioSounds]

wow, this is good,  that is very impressive. When is this going to be commercialised?

[Retroxity]

Enhance! Enhance the enhance! Good. Now just enhance that enhanced enhance. Okay now enhance one more time.

[justinvolmer]

Please use this on the recent Gabby Petito video to see if we can see in the van.

[Jaaaarod]

Oh, ..The books I've read didn't tell what thing like edge detection, Motion flow detect, could to be used like this. You cool!.