Digital Video Tampering Detection TechniquesINTRODUCTION

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(Rafael C. Gonzalez & Richard E. 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 & passive video is known as blind & 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 do 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 passive video spatial tampering detection techniques can be roughly categorized into five category 1) Pixel Based 2) Format based 3) 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 are 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.

BACKGROUND

Video tampering is new in comparison to image tampering. Active and passive image tampering detection play important role to detect tampering in active and passive video. As we have discussed previously that moving images/ frames with time axis is video. So if we want to detect video tampering in passive video or blind passive video then we can take help of passive &blind image tampering detection techniques. Generally MPEG video contain three types frame I) I frame II) P frame III) B frame . I frame is known as Intra frame and have least compression and high quality; P frame is known as predictive frame and have higher compression ratio and less quality in comparison to I frame; B frame is known as bidirectional frame and have highest compression ratio and least quality. I frame of any video is approximately equal to JPEG image. Generally if we want to detect tampering in video we extract frame from video and try to find some clue from that frames .So we can say, indirectly we are utilizing passive image tampering detection techniques in passive video tampering detection techniques. Copy move and splicing is a main images/ frames tampering method. In copy move image/frame tampering some part of image/frame is cloned or copy paste by same image/frame. A lot of copy move detection techniques have been proposed to detect copy move tampering. First solution to this problem is proposed by(J.Fridrich, D. Soukal & J.Lukas,2003) an exhaustive search is performed by comparing the image/frame to every cyclic-shifted versions of itself, which requires (M*N)2 steps for an image/frame sized M by N. They also proposed to use the autocorrelation properties of the image/frame to detect the duplicated regions. Another approach to detect copy-move forgeries is the block-matching method, which divides the image into overlapping blocks. This approach attempts to detect connected image blocks which were duplicated. (A.C.Popesc & H.Farid,2004) proposed PCA(principal component analysis), which work well with additive Gaussian noise and JPEG compression. In the same manner (Li.Guohui, Wu.Qiong, Tu.Dan & Sun.ShaoJie,2007) retrieve features by applying SVD to low frequency wavelet bands. Next challenge for copy move tampering detection is to find out duplicated block in minimum time complexity. Lexico graphical sorting was answer of this problem which sort similar feature vector in minimum time.

However, the computation complexity of the block- matching method could be quite high, and larger resolution make the problem more crucial. (Wang Weihong & Farid H.2007) proposed new computational technique to detect duplication in forge video. Region duplication in frames was detected by matching phase correlation among frames blocks. (Jing Zhang, Zhanlei Feng & Yuting Su,2008) implement region duplicacy concept in static image by introducing wavelet decomposition. (Bo Xu, Guangjie Liu & Yuewei Dai,2012) proposed new idea for fast detection of copy move forgery on the basis of phase correlation. Further, in splicing image /frame tampering we create single image/frame with the help of two frame/image. Several methods have been proposed to detect splicing attack in image/frame. (Y.Q.Shi, C.Chen & W.Chen,2007) proposed a blind splicing detection approach based on a natural image model .The natural image model consists of statistical features including moments of characteristic functions of wavelet sub-bands and Markov transition probabilities of difference 2-D arrays. This method has higher accuracy in comparison method proposed by ( T.T.Ng, S.F.Chang & Q. Sun, 2004). (B. Mahdian & S.Saic, 2009) proposed blind image forgery detection method using noise inconsistencies . This proposed method capable of dividing an investigated image/frame into various partitions with homogenous noise levels, and the detection of various noise levels in an image/frame may signify image/frame forgery. (M.K. Johnson & H.Farid 2005) proposed lighting inconsistence in an image as an evidence of image tampering. This proposed method failed in splicing detection of part with same light direction. (A.C. Popescu & H.Farid,2005) proposed a new method for splicing detection based on CFA Interpolation. (W.Wang, J. Dong, & T.Tan, 2009) proposed splicing detection method based on image chroma. (Xudong Zhao1, Jianhua Li1, Shenghong Li1, & Shilin Wang, 2011) proposed improve method for splicing detection based on Chroma space.

Image tampering detection help in video tampering detection but we should not consider it that we can detect all tampering in video by knowing image tampering detection techniques. Video are very complex in comparison to image(image contain spatial redundancy only but video contain both spatial and temporal) and temporal redundancy play important role in video tampering detection. Intra-frame(Spatial) tampering detection in video is similar to image tampering detection but in interframe tampering detection we utilize the concept of temporal redundancy. Recently (Qiong Dong, Gabo Yang & Ningbo Zhu,2012) proposed MCEA based passive forensic scheme to detect frame based video tampering.

TYPES OF VIDEO TAMPERING ATTACKS

Video tampering attacks are categorized mainly in two categories:

Spatial Tampering

In spatial tampering the content of video frames is modified by the cracker with malicious intension. In spatial tampering there are different method to manipulate the frame like copy move, splicing, resampling etc.

Temporal Tampering

In temporal tampering manipulation is performed on the sequence of frames. Temporal attack-mainly affect the time sequence of the frame. The various temporal tampering attack are as follow, adding new frames, deleting existing frames and shuffling of frames.

Adding New Frame: In this tampering we simply add new frame in the existing video which was not present before to provide fake evidence or for any malicious activity.

Deleting Existing Frame: Intentionally deleting existing frame from the video to remove evidence or any other malicious activity.

Duplication of Frame: In Duplication of frame tampering, people hide the unwanted frame by duplicated frame .

Shuffling of Frame: In shuffling of frame tampering, people change the order of frame which give different meaning than actual.

ISSUE AND CHALLENGES

Video tampering detection is easy if we have active video. Active video contain information like digital watermark and digital signature or combination of both. By extracting digital watermark or digital Signature we can proof the authenticity of video. But if the video is passive and blind which does not contain any digital watermark or digital signature then it will be very tough to detect video tampering. Generally internet streaming videos do not contain the digital signature or digital watermark. Sometime people do not compromise with quality of picture so they do not include digital watermark or digital signature in his camera. So passive & blind video tampering detection is a very challenging area. In surveillance camera we have many similar frames so removing and shuffling frames from surveillance camera are very easy. By removing many frames from surveillance video we can hide useful evidences and information.

The legal implications are important, especially in the case of surveillance video, from which a couple of frames featuring a person walking by could easily be removed. “The question going into court is, How do you prove that it’s really the video that has come

from that particular camera?” We have requirement of advance forensic tool which can prove that this particular video is non-tampered and this particular video is tampered. Proving authenticity of the video is a big issue.

In surveillance systems storage and transmission cost is very important issue. Because surveillance camera recording is going on twenty four hours so there is requirement of huge memory and high transmission speed. In surveillance camera generally background change very slowly and foreground change rapidly so it may be one solution that we record background after some time interval and foreground continuously. But if someone tampered this surveillance recording then it would be very difficult to detect tampering. There are some event based surveillance cameras which start recoding when they observe some event. Tampering detection in event observable surveillance camera is also very tough because they do not include continuous time sequence in the recording. Different type video tampering attack is also big issue for video tampering detection. The various issue and challenge for video tampering detection is as follow:

New Tampering Attack Issue: We have discussed different type of video tampering attack but it not means we have only such type of tampering attack, New tampering attack can arise at any time and tampering detection of such attack is very difficult.

Hardware Issue: Due to the continuous recording in surveillance video it requires huge memory and a lot of transmission power. If we try to make recording event wise then it will be tough to detect tampering in surveillance video because we do not have video frame sequence in continuous time space.

Technical Issue: If we will not have digital signature and digital watermark in the video then proving authenticity of video or tampering detection will be very tough.

Source Issue: If we have no information from which camera this particular video is taken then it create problem to prove authenticity or tampering detection in video.

Tools Issue: video and image editing tool are so advance that they visually left no clue for video / Image tampering detection. These tools are so advance that they create same size frame/ image .Tampering deletion in such frame / image required analysis till pixel level.

Data Set Issue: Video tampering is not as easy as image tampering, so create and maintain a data set of tampered video is very difficult.

New video tampering attack, passive video, low quality video and different format of video are creating great challenge for researcher to detect video tampering .

PROBLEM DEFINITION

In active video we have digital signature and digital water mark to prove authenticity but in passive and blind video we do not have any information like active video. How we shall prove the authenticity of passive video and blind passive video? How we will be able to detect tampering in passive and blind video? This is big question for us.

SOLUTION FOR SPATIAL TAMPERING DETECTION IN PASSIVE VIDEO

According to various method of spatial tampering we can divide spatial video tampering detection technique in five categories. H.Farid(2009).

PIXEL BASED

Cloning/ Copy Move Resampling Splicing

Cloning/ Copy Move

Copy move or cloning is a very popular frame/image tampering method in which we cloned some part of frame/image to conceal some object or person. In copy move tampering we can perform copy move on any portion of image. Generally people perform copy move tampering to hide some evidence.. If we perform copy move tampering carefully then it is impossible to detect it visually. Copy move frame/ image change the original pixel alignment and provide clue for tampering detection. So the solution of copy move tampering detection is to check the pixel alignment in frame/image pixel block.

Figure 1.Copy Move Tampering

Resampling

To create matching composite photograph some time it is necessary to resize, stretch, rotate to portion of frame/image. For example when we are creating composite photograph of two person in which one is fat and another is thin then for making convincing composite photograph we must resize the one person. This process require resampling of previous frame /image in new sampling lattice. Due to new sampling lattice the correlation between neighbor pixel change abnormally which provide clue for tampering detection.

Splicing

Splicing is a most dangerous attack for the frame /image in which we take two frame/image and create one composite frame/image. If we do splicing carefully then it is impossible to detect its joining visually. It is observed that splicing change the higher order Fourier statistics which provide clue for tampering detection.

Figure 2.Spliced frame and original frame

FORMAT BASED

Double MPEG/JPEG Compression JPEG/FRAME Blocking

MPEG Double Compression

If someone want to add or delete or modify the frame/image in existing video then he will first decode it, and after adding, deleting or modifying the frame he will encode it. In this way be get double MPEG compression .This double MPEG compression add special artefact in the video which was not present before. These artefact provide the clue for tampering detection.

Figure 3.GOP of MPEG frame

Frame/Image Block

We know that frame/image is a collection block; block is a collection of macro block and macro block is a collection of pixel. The blocking artefacts use pixel value difference within and across block boundaries. The pixel value difference within the blocks is smaller than the across the blocks. When frame /image is tampered a new set of blocking artefacts may be introduced that do not necessary align with previous block boundaries. These blocking artefacts provide clue for tampering detection.

CAMERA FEATURE BASED

Camera Response Function(CRF) Sensor Noise

Camera Response Function

Because most digital camera sensors are very nearly linear, there should be a linear relationship between the amount of light measured by each sensor element, and the corresponding final pixel value. Most cameras, however, apply a point-wise non-linearity in order to enhance the final image. Differences in the camera response function across the frame/image are then used to detect tampering.

Figure 4.Splicing Detection using CRF (Source: Yu-Feng Hsu, Shih-Fu Chang,2006)

Sensor Noise

Noise produce by Camera sensor in frame/image is known as sensor noise. A digital frame/image 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 . These processing introduces a distinct signature into the frame/image .With the help of nose statistics we can find the clue of frame/image tampering

Figure 5.Gamma correction for better visibility

.

PHYSICS BASED

Light Direction Light Environment based

Light Direction

The falling of light on the surface of object depends on the two thing first one is surface normal and second one is light direction. Measuring the light on different frame we can get clue regarding tampering..

Figure 6. Image or Frame in different light direction ( Source: C. Theobalt, N. Ahmed, E. de Aguiar, G. Ziegler, H. Lensch, M. Magnor, H.-P. Seidel,2005)

Light Environment based

In practice, however, the lighting of a scene can be complex: any number of lights can be placed in any number of positions, creating different lighting environments. Frame captured in different lighting environment have different features. After analyzing these features we can get clue for tampering detection.

GEOMETRIC MEASUREMENT

Principal Point Metric Measurements

Principal Point- In authentic images, the principal point (the projection of the camera centre onto the image plane) is near the centre of the image. When a person or object is translated in the image, the principal points moved proportionally. Differences in the estimated principal point across the frame/image can therefore be used as evidence of tampering.

Figure7.Principal-point-in-camera

(Source:http://www.ds-t.com/software-cd/softwares/Qualup/rhinophoto3d.html-camera)

Metric Measurements

In metric measurement if some part of image /frame is not visible then applying several tools from projective geometry that allow for the rectification of planar surfaces and, under certain conditions, the ability to make real-world measurements from a planar surface. Rectification in planar surface and effect of several different tools provide clue regarding tampering detection.

SOLUTION FOR TEMPORAL TAMPERING DETCTION IN PASSIVE VIDEO

In this tampering detection scheme we are considering MPEG 2 video sequence. MPEG 2 video contains mainly three type of frame I, P, and B. One GOP of MPEG2 video contains 12-15 frames. Starting and ending of frame occur with I frame. I frame approximately is equal to JPEG image. For tampering detection in MPEG 2 video (Qiong Dong, Gabo Yang & Ningbo Zhu.2012.) proposed following process:

1. Extract I,P and B frame from video.2. Extract MCEA value for each P frame.(Motion compensated edge artefacts)3. Calculate MCEA difference of adjacent P frame.4. Apply Fourier transform on this MCEA value.5. If there is any spike then it means video is tampered 6. If there is no spike then it means video is authentic.

RECOMMENDATION

We can apply wavelet transform in place of Fourier transform to get such type of result that after how much time frame tampering occurs. Fourier transform is localize in frequency domain but wavelet is localize in both frequency and time domain. So applying wavelet in this process we can get more precise result related to temporal frame tampering detection.

APPLICATION

Video tampering detection techniques help to prove the authenticity of video & development of computer forensic tool for video. Video tampering detection tools are used by TV Media, Police or some another government or private agencies to check the forgery of video. Everywhere in this world when we have doubt on video we use video tampering detection technique and forensic tools based on these tampering detection techniques.

FUTURE DIRECTION

In future it is going to be a big menace for video security. By analyzing the various video/image tampering detection techniques that were presented in this chapter we can say that the tampering detection techniques are very necessary for the surveillance video, entertainment industry, medical, copyright etc. As the time passes, we are getting more involved with video applications, in our daily lives. Now our information systems most depend on video applications. Various, wide range of tampering attacks, causes severe challenges on information security. In future robustness would be the key point for video tampering detection techniques, so that it can differentiate the acceptable video processing operations from malicious tampering attacks. A perfect passive video tampering detection algorithm that detects all kinds of malicious manipulations and that can tolerate all content preserving manipulations is yet to be discovered. We can hope for the better in the future.

CONCLUSION

Due to availability of lot of video editing s/w it is very easy to tampered video. Tampering detection in active video is easy because we have information like digital signature and digital water mark .By matching digital watermark and digital signature of frame, we can guess which frame is genuine and which is tampered frame. But generally internet video do not contain digital watermark or digital signature so tampering detection in this video is very tough. We have described various tampering detection techniques based on pixel, format, camera, physics and geometry to detect

spatial tampering in passive and blind video/ image. We he also discussed temporal tampering detection process for passive video. Research work on video tampering detection is new in comparison to image, so we are taking advantage of image tampering detection in video tampering detection. Video tampering detection is tough in comparison to image tampering detection. Research work on passive video tampering detection is not complete till now. For getting best and robust passive and blind video tampering detection techniques research work will go on forever.

References:

A.C. Popescu & H. Farid.(2004).Exposing digital forgeries by detecting du- plicated image regions. Dept. Comput. Sci., Dartmouth College, Tech. Rep. TR2004-515.

A.C.Popescu & H.Farid.(2005). Exposing digital forgeries in color filter array interpolated images. IEEE Vol.53. Transactions Signal Processing.(pp.3948–3959).

Baba Mahdian & Stains Lavsac .(2007). Detection of copy–move forgery using a method based on blur moment invariants, Elsevier Vol. 171. Forensic Science International.(pp.180-189).

Bo Xu, Guangjie Liu &Yuewei Dai.(2012). A Fast Image Copy-move Forgery Detection method using Phase Correlation, IEEE Fourth International Conference on Multimedia

Information Networking and Security.china(pp.319-322).

C. Theobalt, N. Ahmed, E. de Aguiar, G. Ziegler, H. Lensch, M. Magnor & H.P. Seidel. (2005). Joint Motion and Reflectance Capture for Creating Relightable 3D Videos. Technical Report MPI-I-2005-4-004.

Li.Guohui ,Wu.Qiong ,Tu.Dan & Sun.ShaoJie .(2007). A sorted neighborhood approach for detecting duplicated regions in image forgeries based on dwt and svd, Proc. ICME, (pp.1750-1753).

Hany Farid .(2009). A Survey of Image Forgery Detection.IEEE Signal Processing Magazine, 26(2) 16-25.

J. Fridrich, D. Soukal, & J. Lukas.(2003). Detection of copy-move forgery in digital images. Proc Digital Forensic Research Workshop, Clevelan, OH.

M.K.Johnson & H. Farid.(2005). Exposing digital forgeries by detecting inconsistencies in lighting.In Proc. ACM 7th Workshop on Multimedia and Security. NewYork(pp.1–10 ).

Jing Zhang, Zhanlei Feng, Yuting Su.(2008). A new approach for detecting copy-move forgery in digital images, published in: IEEE International conference on communication System. Guangzhou .(pp.362-366).

Qiong Dong, Gabo Yang, Ningbo Zhu.(2012). MCEA Based Passive Forensic Scheme for Detecting Frame Based Video Tampering. Elsevier Vol.9.Digital Investigation, Hunan University, China. (pp.151-159).

Rafael C. Gonzalez ,Richard E. Woods . (2002). Digital Image Processing , University of Tennesse,New Jersey 07458: Prentice Hall Upper Saddle River.

T.T.Ng, S.F.Chang, &Q.Sun.(2004). A data set of authentic and spliced image blocks. Tech. Rep., DVMM, Columbia University.

Yu-Feng Hsu & Shih-Fu Chang.(2006).Detecting Image Splicing Using Geometry Invariants And Camera Characteristics Consistency. Paper presented at International Conference on Multimedia and Expo, Toranto, Canada.

Wang W & Farid H.(2007).Exposing digital forgeries in video by detecting duplication. In Proceedings of the Multimedia and Security Workshop, Dallas, TX. (pp. 35-42).

W.Wang, J. Dong, & T.Tan.(2009). Effective image splicing detection based on image chroma. presented in International Conference on Image Processing,Cario, Egypt.

Xudong Zhao1, Jianhua Li1, Shenghong Li1, & Shilin Wang2.(2011). Detecting Digital Image Splicing in Chroma Spaces. Springer IWDW. Verlag ,Berlin Heidelberg. (pp.12-22).

Y.Q.Shi, C.Chen, W.Chen.(2007). A Natural Image Model Approach to Splicing Detection, ACM Multimedia & security. NewYork.(pp.51-62).

Yuting Su, Jing Jhang, jie Liu.(2009). Exposing Digital Video Forgery by Detecting Motion Compensated edge Artefacts, IEEEComputational intelligence and software engineering. Wuhan .( pp.1-4).

ADDITIONAL READING

Athanasios leontaris, Pamel C. Cosman.(2005). Measurement the added high frequency energy in compressed video.in Proc. IEEE International Conference on Image Processig, vol. 2, pp. II- 498-501.

De,A. Chadha, H., Gupta & S.(2006).Detection of forgery in digital video. In Proceedings of the 10th World Multi-Conference on Systemics: Cybernetics and Informatics,2006, vol. V, pp. 229-233.

Hsu,C. Hung,T. Lin,C., & Hsu,C .(2008).Video forgery detection using correlation of noise residue. In Proceedings of IEEE Workshop Multimedia Signal Processing (MMSP), 2008, (Cairns, Queensland, Australia, Oct.2008), pp. 170-174.

H. Farid.(2006). Digital doctoring: how to tell the real from the fake .Significance, 2006, 3(4): 162-166.

Kobayashi, M. Okabe,T., & Sa to,Y. (2009). Detecting video forgeries based on noise characteristics. In Proceedings of the 3rd Pacific-Rim Symposium on Image and Video Technology, Tokyo, Japan, LNCS 5414, 2009, pp. 306-317.

T. Sikora.(1997).Digital Consumer Electronics Handbook. chapter MPEG-1 and MPEG-2 Digital Video Coding Standards.

Wang W & Farid H.(2006) .Exposing digital forgeries in video by detecting double MPEG Compression. In Proceedings of the Multimedia and Security Workshop, Geneva, Switzerland, pp. 37-47.

Wang W & Farid H.(2007).Exposing digital forgeries in interlaced and de-Interlaced Video. IEEE Transactions on Information Forensics and Security, 2007, vol. 2, no.3, pp.438-449.

Wang W & Farid H.(2007).Exposing digital forgeries in video by detecting duplication . In Proceedings of the Multimedia and Security Workshop, Dallas, TX, 2007. pp. 35-42.

Weihong Wang, Hany Farnd, Exposing Digital Forgeries in Video by Detecting double MPEG compression, 11th ACM multimedia and securit workshop Princeton, NJ, SEP 07-08, 2009, pp.39-47

Weiqi Luo, Min Wu, Jiwu Huang, MPEG recompression detection based on block artifacts, Proceedings of the SPIE on Security, Forensics, Steganography and Watermarking of Multimedia, Vol.6819, pp. 68190X-68190X-12, 2008.

KEY TERMS AND DEFINITION

ACTIVE VIDEO: A video which contain digital signature and digital watermark is called active video.

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

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

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.

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

CRF: The camera response function measure 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 .

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 .

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

GOP: Group of pictures is 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.