|J Pathol Inform 2013,
Review of "Digital image forensics: There is more to a picture than meets the eye" by Husrev Taha Sencar and Nasir Memon (Editors)
Department of Pathology, University of Illinois at Chicago, 840 South Wood Street (130 CSN), Chicago, IL 60022, USA
|Date of Submission||09-May-2013|
|Date of Acceptance||14-May-2013|
|Date of Web Publication||29-Jun-2013|
Department of Pathology, University of Illinois at Chicago, 840 South Wood Street (130 CSN), Chicago, IL 60022
Source of Support: None, Conflict of Interest: None
|How to cite this article:|
Levy B. Review of "Digital image forensics: There is more to a picture than meets the eye" by Husrev Taha Sencar and Nasir Memon (Editors). J Pathol Inform 2013;4:17
Sencar, HT, Memon N. (Editors)
Digital Image Forensics: There is More to a Picture than Meets the Eye. Springer: New York; 2013.
Photography has existed for over a century, creating visual records that have been used extensively in many different ways, including for our legal system. Traditional film photography has always been susceptible to doctoring, but it required persons with specials skills making a significant effort for each altered image. This scenario has been turned upside down with the advent of digital imaging. Tools that allow the simple and quick modification of digital images are available to everyone. Most of us have used these tools to crop an image, eliminate "red-eye" or otherwise innocently improve an image. Unfortunately, these same tools can also be used to doctor images with nefarious intent, creating an image that is not a representation of reality. This is a challenge in many areas, but none with more serious consequences than our legal system.
The field of digital image forensics has recently emerged to scientifically evaluate these images and provide reliable answers regarding the origin and truthfulness of digital images. Experts in this field may be called on to locate images that may have been deleted, confirm that obtained images came from a particular device at a specific time, analyze images for alterations, authenticate the image and ultimately testify to this in a court of law. These tasks are made more difficult by the rapid evolution of the technology and the development of techniques to counter the forensic analysis of images.
"Digital Image Forensics" was written to provide a comprehensive overview and understanding of this field and it is quite successful at achieving this goal. However, it is written for an audience of persons with expertise in imaging and digital imaging. The level of mathematics and theory used in this text is beyond the ability of laypersons or most people outside this field. The authors have included some real world examples within the text that anyone could utilize to understand the depth and scope of the subject matter. The text is well-organized. Each chapter contains an abstract, an introduction with a brief outline of the material presented, well-defined sections within the body of the chapter, summary and ample references.
The first part of the book, consisting of four chapters, provides a background on digital images that allows the reader to later understand how forensics can identify the source of the image and detect doctoring of these images. The authors first introduce the formation of the digital image and the camera hardware that includes the lens, various filters and the sensor. Both charge-coupled devices and complementary metal oxide semiconductor sensors and artifacts that these sensors can produce are described. The second chapter addresses how digital cameras have multi-step processes to transform the raw image through a variety of mechanisms and techniques into the "finished image." They note how cameras and manufacturers differ in their application of this processing. The third chapter describes the various formats that images are stored and may be compressed, starting with a description of the different coding techniques that can be used. Both lossless compression (graphic interchange format, portable network graphics and tag interchange file format) and lossy compression (Transform Coding, joint photographic experts group [JPEG] and JPEG-2000) are described. A description of how metadata is stored is also included. The last chapter in this part describes how to search for and extract digital images from devices. They describe how files are stored and indexed in both the file allocation table and new technology file system. Techniques for extracting "deleted" files are presented with some real examples.
The second part of the book focuses on the current techniques for identifying the source of a given image. The first of three chapters in this part focuses on the identification of the class of the device that created the image. Studying the specific techniques applied to process an image can lead the forensic expert to identify the particular brand and model of device that captured the image. The next two chapters in this section study the defects in sensors that can provide a "fingerprint" of an individual device, allowing the forensic expert to conclusively tie an image and device together, determine the temporal sequence of images and identify images that have been altered. They describe not only sensor defects as photo-response non-uniformity, dark current and pixel output, but also scratches and dust present on the devices.
The third part of the book, consisting of three chapters, describes and demonstrates the techniques for verifying the integrity and authenticity of image evidence. The authors first introduce the topic of natural image statistics and how by understanding natural images the forensic expert can detect photorealistic (or computer synthesized) images, steganalysis (the inclusion of hidden messages in an image) and rebroadcast image detection (taking an image of an image). The second chapter describes in more detail the currently available tools for detecting doctored images and includes several examples of how these tools function. The third chapter focuses on the topic of computer synthesized and recaptured images. This is an important topic in the forensic world because while it is illegal to possess child pornography, it is currently not illegal if the images are not real, but computer synthesized (photorealistic).
The fourth and final section of the book addresses some of the practical considerations in this field. A chapter involving courtroom considerations for digital image forensics speaks to some of the legal issues and is probably the easiest chapter for the average reader to understand. The final chapter of the book focuses on the fascinating area of counter-forensics, which is how people are developing techniques to attack the work of image forensics. Images can be altered to change the device fingerprint, making it appear as if another device obtained the image. Techniques can be used to hide evidence of tampering from commonly used detection methods and suggestions for detecting these techniques are mentioned. This is clearly the cutting edge of this field today.
In summary, this is an excellent book for someone interested in the theory and application of digital image forensics that already has a background in mathematics and the physics of imaging. It is a thorough presentation of the current state-of-the-art. One example where they demonstrated that the cover of a tabloid newspaper was a doctored image was revealing. After reading this book, you may never accept an image at face value again. There is truly more to a picture than meets the eye.