Holmes: An Open-source Image Forensics Toolkit

ABSTRACT

 

Digital images continue to be ever-present in our daily lives as means of conveying information. With the availability of image editing software, now more than ever our ability to trust the images we see have been brought into question; the field of digital image forensics aims to solve the problem of image authentication as well as verifying the integrity of images. Currently, breakthroughs in the field are locked behind proprietary software and opaque research papers. This makes the fight against misinformation significantly harder. This project is based on the idea that a transparent collaborative effort in the field of digital image forensics is required if society is to ever hope of curbing the problem of misinformation propagated with digital images.

CHAPTER 1: INTRODUCTION

1.1     Overview

The digital age is one where we get constantly exposed to a vast array of digital visual media. Due to rapid developments in computer technology and the internet; images, and videos, today have become the common source of news and facts, including daily controversies, trials, and surveillance. With the advent of low-cost cameras, rapid developments of easy-to-use image-editing software like Adobe Photoshop, GNU’s Gimp, and other digital doctoring tools are more available to ordinary users. However, our understanding of the ethical and legal implications associated with doctoring images remains limited. When does an innocent crop or minor filtering become a legal threat to society; this is a question that continues to go unanswered. Because of the easy-to-tamper nature of digital media like images, there is an urgent need to improve our ability to effectively and efficiently detect any modification that an image has undergone.

Image tampering is not a new phenomenon; not long after Niepce created the first photograph in 1814, had photographs were already being manipulated. Stalin, Hitler, and many others had photos doctored to rewrite history and distort people’s perception of reality. Men like these understood the power images had; back then, image manipulation required a combination of cumbersome and time-consuming darkroom techniques. Today, technologies like Adobe Photoshop and other similar software have made image fakery easier while making the resulting fake images are often hard to detect. The combination of these faked images being easy to produce and difficult to identify has a significant impact on society. Doctored images now appear on mainstream media outlets, social media, dating websites, political campaigns, and even scientific journals. These technologies for manipulating images are developing at headlong speed while developments of technologies required to detect such manipulations desperately need to catch up.

Digital image forensics (DIF) research has the task of analyzing and unearthing facts about images. Which involves image source identification, image tampering detection, and hidden data recovery with no prior registration of the digital image. This new discipline stems from existing security-related research that exploits image processing and analysis techniques to recover information about a digital image. The main research paths include:

1. ballistic analysisor image source identification: which aims to identify the device which captured a given image or at least determine if an image wasn’t created by a particular device.
2. image tamper-detection: This aims to expose traces of manipulation by studying the nature of natural images.

The goal of this project is to create an easily accessible and extensible application for digital image forensics, which serves as a starting point bundled with popular image forensics tools and focused on being extensible, making it easier to experiment with algorithms found in the latest research on DIF.

1.2     Statement of the Problem

Digital image forensics(DIF) is still an emerging research field that is attracting growing interest from the scientific community over the last couple of decades, with important implications for ensuring the credibility of the content we consume.

With a growing number of publications in the field, there is a need for an easily accessible and extensible environment for the various DIF algorithms found in the latest workshops and research papers. Currently, the field of DIF relies mainly on security-by-obscurity; current solutions come with high retail prices and are reserved only for law enforcement and government agencies. This model is not sustainable as the barrier to entry for image manipulation is low while the barrier for entry into DIF is high, which is significantly slowing down progress in the field.

 

1.3     Aims and objectives

The main aims and objectives of this project are as follows:

1. Allow users the ability to perform image analysis like metadata extraction, thumbnail extraction, etc.
2. Implement a popular and effective image tamper detection tool like Noiseprint.
3. Allow users to compare two images.
4. Generate a report that could be exportable as a PDF or printable format.
5. Create an index of quantization tables for classifying images and matching images to their possible device of origin.

 

1.4     Scope of the Project

 

This project does not aim to produce an automatic solution to the problem of digital image forgery. At the end of this project, it is expected that Holmes can serve as a home for digital image forensics tools and breakthroughs in the field of digital image forensics.

The chapters of this documentation will be organized as follows:

• Chapter 1: Gives a general overview of the whole project, focusing on the significance of digital images and the rise of image forgery.
• Chapter 2: Provides a review of existing literature on the subject of digital image forensics, including popular methods of detecting forgeries.
• Chapter 3: describes the project's requirements, analysis, and design.
• Chapter 4: includes the implementation of the project and issues faced.
• Chapter 5: provides a conclusion, future additions, and recommendations for the project.

Holmes: An Open-source Image Forensics Toolkit

Sharing is caring!