Digital Video Tampering Detection Techniques

Digital Video Tampering Detection Techniques

Ramesh ChandPandey (Department of Computer Engineering, Indian Institute of Technology (BHU), India), Sanjay Kumar Singh (Department of Computer Engineering, Indian Institute of Technology (BHU), India) and K.K. Shukla (Department of Computer Engineering, Indian Institute of Technology (BHU), India)
Copyright: © 2015 |Pages: 11
DOI: 10.4018/978-1-4666-5888-2.ch125
OnDemand PDF Download:
$30.00
List Price: $37.50

Chapter Preview

Top

Introduction

We can define video many ways, Time varying image is known as video or Changing of Image in temporal domain is known as video or Transformation of 4 D(X,Y,Z,T) physical object in 3 D(X,Y,T) is known as video.(T-temporal domain, (X,Y,Z)- Spatial Domain).An image is defined by spatial coordinates(X,Y)and its intensity function F(X,Y). When (X,Y) and intensity value is discrete at every point in image plain then we call image digital image(Gonzalez & Woods,2002). Due to high availability of low cost s/w editing tools, it is very easy to tamper the digital video. Some modification in video does not lead to malicious tampering in video for example modification in video to increase quality of video. Illegal, improper and malicious intension for modifying video to conceal some important information, event or object is known as video tampering. According to video we can divide video tampering detection techniques in two categories: first one is active video tampering detection techniques and second one is passive video tampering detection techniques. In active video tampering detection techniques we use the concept of digital Signature and digital watermark or combination of both. But in passive video tampering detection techniques we do not have any information regarding digital signature and digital watermark. If we have no information of camera from which video was taken then we call it blind video. The techniques used to detect tampering in blind and passive video is known as blind and passive video tampering detection techniques. Video tampering and tampering detection both are tough in comparison to image tampering and tampering detection. People follow the concept seeing is believing but video tampering has disproven this concept. Video tampering detection is necessary because people are using video tampering to defame popular person, concealing important information and presenting it as proof in the court to get judgment in his favors. If we have active video then it is easy to detect tampering by using digital signature and digital watermark, but if we have no information about source camera and video does not contain digital signature or digital watermark then it is very challenging to detect video tampering. Generally Internet streaming video does not contain information regarding source camera, digital signature and digital watermark. Blind and passive video tampering detection is new era for researcher and research work in this area is going on. In video mainly three types tampering arise first one is spatial tampering second one is temporal tampering and last is spatial-temporal tampering. In spatial tampering we generally focus on intra frame but in temporal and spatiotemporal we focus on interframe. In passive video spatial tampering detection techniques can be roughly categorized into five category 1) Pixel Based 2) Format based3) H/W or Camera based 4) Physics based 5) Geometric based H.Farid(2009).In pixel based tampering detection we mainly focus on intra frame and spatial coordinate of intra frame. The various video tampering approach in this category is Copy Move, Splicing, and Resampling. Format based technique include Double MPEG compression, MPEG Blocking etc. H/W or Camera based tampering detection use Sensor Noise, Color filter array, Camera response function, Chromatic aberration, White balancing and gamma correction features of Camera used in shooting video. In physics based video tampering detection we mainly focus on light direction and light environment for video tampering detection. In geometric based tampering detection we mainly focus on principal point and Metric measurement. If we want to detect temporal tampering in passive video then we can use the concept of motion compensated edge artefacts (MCEA) for I, P and B frames in video.

Key Terms in this Chapter

SVM: Support vector machine.

Histogram Oriented Gradient: ( HOG): Used to find local features of video frame.

FMT: Fourier Mellin transform.

Scale In Variant Feature Transform: ( SIFT): Used to find local features of video frame.

Blind Video: A video which does not contain any information regarding source from which these video was taken then this is called blind video.

Sensor Noise: Noise produce by Camera sensor in image is known as sensor noise. A digital image/frame moves from the camera sensor to the computer memory, it undergoes a series of processing’s, Including: quantization, white balancing, De-mosaicking, color correction, gamma correction, these produce noise in image.

Color Filter Array: ( CFA): used for filtering the color which is measured by camera sensor.Generally we see three color filter in camera, red green and blue.

Group Of Pictures: ( GOP): A collection of 12-15 frame of video. According to format of video GOP contain different type of frame. MPEG 2 video contain three type of frame I, P and B. Starting and ending of GOP occur with I frame.

Speed-Up Robust Feature: ( SURF): Used to find local features of video frame.

Demosaicing: Demosaicing is a process to find out missing color from existing color sample with the help of Color filter array. Other nameof demosaicingare color reconstruction or color filter array interpolation.

Principal Component Analysis: ( PCA): Used to find global feature from image /frame.

Splicing: Splicing is an image or frame tampering technique in which we create one image or frame with the help of two image or frame.

Camera Response Function (CRF): measures image/frame irradiance at the image/frame plane to the measured intensity values. Various application like Color constancy, photometric stereo, and shape from shading, require object radiance rather than image/frame intensity.

Active Video: A video which contain digital signature and digital watermark is called active video.

Copy Move: Copy Move or cloning is an image or frame tampering technique in which we remove scene or object within that particular image or frame.

Passive Video: A video which does not contain digital signature and digital watermark is called Passive video.

Complete Chapter List

Search this Book:
Reset