Clones Blocks and Clones Keypoints: which one is better?

Clone detection (aka “copy-move detection”) is a very important image authentication task. Clones are a special case of image manipulation where part of an image is copied, possibly resized, rotated, sheared, etc., and then pasted to another region of the same image. The two main applications of cloning are:

  • creating multiple (fake) copies of an object through copy-paste;
  • removing an object from the scene by covering it with a cloned portion of the background.

This is explained with a very simple example in the image below.

Two possible ways of using copy-move to create a fake image.

The image forensics research community worked hard to develop techniques for clone detection, and two main approaches have been invented: block-matching and keypoint-matching. As suggested by their names, they are based on two different strategies, briefly explained below.

Block-matching approach

  1. Split the image in overlapping blocks;
  2. Compute a digest (“descriptor”) for each block, possibly robust to rotation, scaling, compression, etc.;
  3. Search for clusters of matching descriptors.

Keypoint-matching approach

  1. Detect keypoints (SIFT, SURF, BRISK, etc.) from the image;
  2. Compute keypoint local descriptors;
  3. Search for (clusters of) matching keypoints.

Which one is better? It depends, and we try to explain why with the table below:

So, if your question was: “Do I need a block- or a keypoint– based algorithm for my analysis?”, the answer is: you need both!

That’s why Amped Authenticate features both algorithms under the Local Analysis category: Clones Keypoints and Clones Blocks. Let’s compare their output on the sample image we used in this article:

We see that the cloned seagull (top row) is detected by the Clones Keypoints despite the strong down-scaling applied to the cloned object; such a geometrical transformation is too strong to be detected by Clones Blocks. On the other hand, Clones Blocks successfully detects the cloned background (bottom row), that is not detected by Clones Keypoints because the cloned area is just too flat and it does not contain enough keypoints.

We hope you enjoyed this quick tip! Stay tuned and don’t miss our next #ampedtiptuesday post!

Introducing our New Tip Tuesday Series

Happy New Year – this is our first post of 2019!

We are present in many industry and partner events worldwide, but unfortunately, we can’t always meet our global customers face to face to show you how to get the most out of your Amped solutions. So we’ve created a new blog series called Tip Tuesday (#ampedtiptuesday). We hope this year will be full of useful tips and tricks for all our solutions.

We will teach you things in Amped FIVE like how to organize and customize the panels, how to use the Assistant, how to use the concatenation in Covert DVR, and the Screen Capture tool.

For Amped Authenticate, you will learn about differences between Clones Blocks and Clones Keypoints, how to find reference images for your case, how to use PRNU on images gathered from social media, and how some minor metadata may tell more than it seems.

And much much more!

If you have suggestions for things you would like to learn about, send them our way!

Our first tip will be published next Tuesday – don’t miss it!

Amped Authenticate Update 12336: Brand new smart report tool, PRNU detection improvements and more…

The festive season is right around the corner and this is one of our busiest times of the year! Despite this, we’re here with another update to Amped Authenticate just in time for the holidays!

Smart Report

While we’ve included the ability to generate batch processing reports for a while, we’re now introducing an exciting new reporting method designed to make it quicker and easier to report relevant filters.

Continue reading

DVRConv Update 12151: Admin lock down settings ability and more formats now supported

Yes, you heard correctly – we’re here with yet another update! We’ve added some highly sought-after features to Amped DVRConv, the fastest and easiest way to convert proprietary video formats.

Settings Lock Down Option

Sometimes system administrators need to lock down the settings of Amped DVRConv in order to avoid having users modify the settings accidentally, which could disrupt the workflow of the lab; for example setting a conversion to transcode instead of stream copy. Now it is possible to protect these settings, allowing only system administrators to modify them!

Continue reading

2018 LEVA DME Training Symposium – Amped Review

Hi everyone, great to be back writing a quick blog post on here. We are SO busy here at Amped with so much going on… but I do like to chat with you all when I can. 

This year’s LEVA event was in San Antonio, Texas.

What a place! If you ever get the chance to get down there, it’s a wonderful city.  

During the event, I held a 12-hour workshop on using Amped FIVE. Many people had received some training before, but some were new users, and some didn’t even have the software yet. That’s one of the many things I love about FIVE — it’s so quick to get using, once you understand the interface. 

The workshop concentrated on task and question orientated workflows —  how to complete a desired task, and how to answer a specific question. The magic of FIVE is that you can do multiples of these in a single project. 

No need to reformat or transcode in one piece of software, then edit images in another, then process video in one more. Keeping things simple means a much quicker result and that’s what we are here — to help you with getting results!

We looked at many functions and filters so I thought I would give a rundown here. It may help as a review to those who attended or be of interest to all you guys out there who may have missed a blog post over the past few years. 

Please note though that workshops like these allow you to see just some specific features of Amped FIVE — you really should attend the official Amped FIVE training to learn about all the possibilities in FIVE. Check the schedule here.

Continue reading

Perspective Stabilization and Perspective Super Resolution

In our latest update to Amped FIVE, you’ll find two new filters that work together to stabilize and enhance video with an object that has some change in perspective as it moves: Perspective Stabilization and Perspective Super Resolution.

Let’s take a look at how they work!

You can also watch the two filters in action in our latest video found here:

Continue reading

The International Meeting of Criminology and Criminalistics

The Forensics Meeting is an international event dedicated to the various areas of forensic science, organized by the Institute of Forensic Sciences (ISF: Istituto di Scienze Forensi) in collaboration with the National Association of Criminologists and Criminalists (Associazione Nazionale Criminologi e Criminalisti).

The event will be held on November 9-10, 2018 in Milan, Italy, and is open to professionals, professors, and researchers working in the field who want to deepen their knowledge and get updates on the latest news and connect with international experts and companies.

Martino Jerian, CEO and Founder of Amped Software, will explain some of the potential problems when investigating image and video evidence and how to solve some technical issues in various potential scenarios related to CCTV and images.
Martino’s presentation entitled Forensic Video Analysis will be on Friday, November 9 from 15:15-16:00. 

For more info and to register visit: https://www.forensicsmeeting.com/

Amped FIVE Update 12076: Automatic Perspective Stabilization for License Plates and much more!

The festive season is on its way, but Team Amped are here with an early gift! We have another update to Amped FIVE and this time it’s huge!

Perspective Stabilization and Perspective Super Resolution

We’ve now created two new filters that work together to stabilize and enhance video that includes an object with a change in perspective as it moves. The most common scenario for this is the need to enhance the license plate of a moving car filmed by a static CCTV. While it has been possible to correct perspective issues since early versions of Amped FIVE using Perspective Registration, this could often be very time consuming and not always easy. With the Perspective Stabilization this step has now been basically automated. Of course, Perspective Registration is still there for you for the extra-difficult cases and for manual selection.

We’ve dedicated an entire post to Perspective Stabilization and Perspective Super Resolution which will be published in the next few days.

Replace Channel

Another new filter! Replace Channel replaces one color channel with another. It’s found under the Channels filter set.

Replace Channel converts the image into another selected color space and replaces one of the channels with another. Let’s take this image, for instance:

Continue reading

LEVA 2018 Annual Digital Multimedia Evidence Training Symposium

LEVA’s 29th annual video and audio evidence symposium will be held in San Antonio, Texas, from November 5-9. This annual training symposium provides valuable workshops and lectures to elevate skills set and offers great networking opportunities. LEVA will direct its attention towards addressing critical issues for those involved in the retrieval, processing and presenting in court all types of digital multimedia evidence.

Visit Amped Software during the vendor day on Tuesday, November 6 from 9 am – 5 pm to see first hand why Amped Software products are now the global leading forensic image and video solutions for law enforcement agencies worldwide.

Also, don’t miss the Amped FIVE workshop entitled: One Tool…Amped FIVE…Countless Possibilities! presented by David Spreadborough, Amped’s International Trainer.

Over the last few years, Amped FIVE has fast become an indispensable weapon in the FVA arsenal. It doesn’t matter if you are a daily user, or have never used it at all; if you’re interested in the latest developments on decoding, workflow, restoration, enhancement, comparison, and presentation – this workshop is for you! See how using FIVE can expedite your investigations, using a scientific methodology to get results – fast!

The workshop is divided into two parts: 
Part 1 on Thursday, November 8, from 8 am – 5 pm
Part 2 on Friday, November 9, from 8 am – 5 pm 

Established in 1989, LEVA is a 501(c)(3) nonprofit corporation committed to provide advanced training and certification in the science of forensic video analysis. LEVA serves as a key resource providing opportunities for professional development through quality training and informational exchange.

For more info visit: https://leva.org/index.php/training/training-conference