1.1Introduction
A digital image is the representation of a real image, being constituted as a two-dimensional data structure, which can be natural numbers, complex numbers, or real numbers represented on a finite number of bits. An image element is called pixel, the whole image being composed of a string of octets. The color components of pixel are red, green, and blue (RGB). The image formats are known according to their file extension, such as bitmap (BMP), the gross form of the image; Joint Photographic Expert Group (JPG), intended to compress the images taken from reality; Graphic Interchange Format (GIF), a good quality format and very strong compression; Tagged Image File Format (TIF), used for professional photos due to its versatility and nondestructive compression; Digital Imaging and Communications in Medicine (DICOM), used for medical images; and raw, which contain the entire information that was captured by the sensor at the moment of the shooting.
In real and legal practice, there are situations where fraudulent images are used. Detecting a fake ordinary image or video edits is not an easy task. Until recently, only few software could determine whether an image was original or whether someone manipulated it. Currently, the image can be cracked by using a technique called Deep Video Portraits (DVP) [1], which allows the imitation of a great variety of movements, including a three-dimensional (3D) head view, its rotation, shifting it, and even blinking. This uses artificial intelligence by which movement data are taken from the source and transmitted to the digital image that is to be integrated into the video. Through artificial intelligence, motion patterns are collected, analyzed, calculated, and predicted in order to create a more realistic video [2]. DVP is a very effective method for creating computerized animations, converting into realistic photo images, and is extremely easy to confound with a person’s real image.
Some software that implement many filters that work automatically or semiautomatically are already available. Software developers have created 3D reconstruction algorithms over 50 smart filters, most of which are unique. These algorithms are part of the specialized software, and some of them are embedded into the visual video cameras. There are companies that patent at least five new technologies related to photo and video processing. Among these, there are algorithms that allow to perform 3D face and object reconstruction, measure the dimensions and distances between objects on photo and video, analyze the motion model of an object in a frame, etc. VOCORD VideoExpert is a powerful tool used for analyzing and improving digital photo and video materials [3]. The program restores indescribable objects on the original photos or videos – faces of people, license plates, or text, greatly improving the quality of the material.
Amped Authenticate is another software package used to authenticate forensic images and the incorrect detection of digital photos [4]. Current products and processes focus on one or several scientific instruments. With the help of authentication, Amped empowers law enforcement experts and legal medicine laboratories with an instrument capable of detecting any manipulation that has been done on an image and checking whether a digital photograph has been generated by a specific device. Previously, specialized applications have been highlighted in this domain, but there are other applications that can perform a similar analysis of the image (satellite), the best known being Peripheral Component Interconnect Geomatica [5], Environmental for Visualizing Images (ENVI), or ER Mapper [6].
Exelis and Environmental Systems Research Institute have joined to improve the interoperability between image processing and geographic information system (GIS) software through ENVI and ArcGIS. Practically, ENVI is a selection of workflows and tools that have been integrated into the ArcGIS ArcToolbox environment.
Most existing applications in this domain are professional applications that offer, in most cases, the “package” functionality at a fairly high cost. As a rule, they are marketed as a trial version, valid for a limited time, or as a simple editor or viewer that does not include any functionality. In addition to this impediment, many of them come as third parties, which require integration with other applications. An example is ENVI. If it is desired to use some of the processing or methods that it offers, another application, such as ArcGIS, which may be available at a cost, is required in order to enable the use of the ENVI functionalities.
These applications may also require a user guide or course as their complexity makes them inaccessible for some users. Few of them even require an Internet connection for certain packages and functionalities. Our software was developed in order to avoid these drawbacks of the existing applications and make it accessible to everyone.
This chapter is organized as follows: the second section presents some techniques related to image processing, the third section is about the theoretical background, the fourth section presents the application and its implementation, and the fifth section evaluates and validates it. Finally, few conclusions are presented in the conclusion section.
