Human wrote:Hello Equinox
I have no idea how to capture a frame and post it here or create arrows to show exactly what I mean, if any of you can tell me how to do this it would be greatly appreciated and would help me in the future to show things better, maybe someone can tell me how in a private message so as not to stear away from these post's?
Equinox wrote:Here I will explain it easy for you how to do it because after all, You are "only Human" Oh I made funny!
Thank you very much for the tool tips and info on how to use them, I will download them and learn how to use them in order to submit more here, very appreciated.
fbenario wrote:EDIT: Equinox, many thanks for your instructions on all those tools. I'm saving them.
EDIT NO. 2:
Admins, please pin a copy Equinox's post at the top of the Tools Of The Trade thread, so it always appears at the top of that thread. Each member needs easy continuing access to this user-manual for photo-analysis. Thank you very much!
Keeping Close Track of Chats, Word for Word
People store their emails, photos and documents online to keep them from being lost or accidentally deleted. But what about the records we never save to begin with, like phone conversations and text messages? These hold a lot of useful data and can sometimes be the only point of reference for important conversations.
Calltrunk is a service that records, stores and transcribes phone calls while Uppidy does much the same thing with text messages. WSJ's Katherine Boehret says they're both great ideas but says there might be privacy concerns.
I've been testing Calltrunk, a service that records, stores and transcribes calls initiated by its app or website; a manual feature on iPhones and Skype also enables recording incoming calls.
reichstag fireman wrote:Discrete Wavelet Transforms
Neil Krawetz describes in his excellent 2008 paper "A Picture's Worth: Digital Image Analysis and Forensics" the use of the Wavelet Transformation to detect image manipulation.
Krawetz briefly explains the science behind it: a suspect image is re-encoded using wavelets to approximate the image (aka the signal). The re-encode is repeated, but in each re-encode, an increasing percentage of available wavelets is used. A gallery of re-encodes of the original image is created, with each re-encode containing a different count of wavelets. Increasing the count of wavelets increases the sharpness and colour accuracy in the re-encoded image.
This forensic technique exploits a property of the wavelet transform: as the wavelet count increases, the components of a composite image tend to sharpen at different rates. This is the key to detecting photo fraud using the DWT.
The DWT can sometimes detect image fraud where an analysis of the error layer has failed to uncover anomalies. What Krawetz doesn't describe is the ease with which an amateur can implement the wavelet transform technique for image forensics.
The standard JPEG (1992) image standard utilises a Discrete Cosine Transform (DCT) to shift the energy from the spatial domain to the frequency domain for compression. However, the little-used JPEG2000 (jp2) standard uses a Discrete Wavelet Transform instead. Tools for processing the jp2 standard can be used for this wavelet-based forensic technique.
The image rendering engines in most web browsers and many graphic tools cannot natively handle jp2 images, but the Open Source ImageMagick software suite has a range of tools and library support for jp2. The ImageMagick website documents wavelet-based encoding using the JPEG2000 image standard: http://www.imagemagick.org/script/jp2.php
Below is an example of an image composition of a vicsim.
ELA@0.70 highlights the image manipulation and this is confirmed by the DWT forensic technique.
Error Level Analysis (ELA)
Discrete Wavelet Transform (DWT) Analysis
Notice how the neck remains fuzzy yet the necklace soon comes into focus as the utilisation of available wavelets is increased. This is compelling evidence of image manipulation.
Users browsing this forum: No registered users and 2 guests