Identify Photoshopped Pictures [ Analysis ]

Identify Photoshopped Pictures [ Analysis ]

Adobe Photoshop is a very creative software that enables too many features and flexibility, which gives users  the ability to make creative and magical pictures with it. But some people uses Photoshop to create fake images, re-touched pictures, color correction, fake effects... and more, while it's hard to identify using the human eyes. Photoshop professionals can make everything believable using their magical touch, but you can easily identify if a picture is Photoshopped or not using this non-human -robot- website identification.

 How it Works ?

This web based app identifies the errors that can produce an image using the Image Error Level Analysis which is used to highlight the compressed pixels after extracting a heat map out of a JPEG image, since a Photoshopped image is saved not once but several times over and over, then it identifies the compression that was generated by those saves during Photoshopping. After highlighting those important pixels, it creates some extra layers 
(colored or black & white) to identify the Photoshop edits that have been made to that picture. And finally it gives a simple result telling you if it's Photoshopped or not and the effects applied to that image.

Does it Identify only Images by Photoshop?

No this web app is designed to identify any sort of modifications that happened to an image from any image editing apps, it surely identifies Gimp images, Pixlr images, Photo Effects pictures and more...

How to Identify Photoshopped Pictures

1) Go to PSKiller web app

2) You have the option to Choose an Image from Computer or From URL

3) Hit Submit. Wait few moments for uploading the picture and analysis time.

4) The Analysis result should be showed in the bottom of the picture.

5) It will shows up the different effects and mods that have been applied to the photo submitted. I've edited Mark Zuck's picture by changing: Brightness and Contrast, Color balance and saved it twice so the compression result is twice as it shows in the above image. There's different results according and depending of the picture submitted.

Note: Sometimes it fails to identify Photoshopped pictures because it's based on some algorithms that are not very smart but uses some advanced techniques like  Image Error Level Analysis.