The sequence of processing steps are performed during multimedia acquisition are shown in the Fig. 1.1.
Fig 1.1 Sequence of steps during digital multimedia acquisition [1]
Firstly, a digital camera takes the light rays and focuses it onto imaging sensor with the help of lenses. An imaging sensor consists of an array of photodiodes. When light strikes this array, each photodiode generates
an analog signal which is then converted to digital signal by built-in analog to digital convertor. Most of digital camera manufacturers use a charge-coupled device (CCD) as the image sensor. Now a day CMOS chips are also emerging as a popular alternative. These photodiodes just record the brightness of light i.e. they are not sensitive to colors and hence produce a monochromatic output. So in order to produce a color image, a color filter array (CFA) is used before the imaging sensors. CFA is a thin film on the sensor that passes a certain color component of light through it. Therefore, each photodiode records the light intensity for a single color only. This CFA pattern varies from one manufacturer to another. Fig.1.2 shows the commonly used Bayer CFA pattern. As a result the raw image generated by imaging sensors consists of red, green and blue (RGB) intensity values as shown in Fig. 1.2. To obtain the color image, the obtained RGB intensity values need to be interpolated. In order to perform this interpolation, demosaicing algorithms are applied which estimate the missing intensity values based on the values of existing neighbors in each of red blue and green channel.
an analog signal which is then converted to digital signal by built-in analog to digital convertor. Most of digital camera manufacturers use a charge-coupled device (CCD) as the image sensor. Now a day CMOS chips are also emerging as a popular alternative. These photodiodes just record the brightness of light i.e. they are not sensitive to colors and hence produce a monochromatic output. So in order to produce a color image, a color filter array (CFA) is used before the imaging sensors. CFA is a thin film on the sensor that passes a certain color component of light through it. Therefore, each photodiode records the light intensity for a single color only. This CFA pattern varies from one manufacturer to another. Fig.1.2 shows the commonly used Bayer CFA pattern. As a result the raw image generated by imaging sensors consists of red, green and blue (RGB) intensity values as shown in Fig. 1.2. To obtain the color image, the obtained RGB intensity values need to be interpolated. In order to perform this interpolation, demosaicing algorithms are applied which estimate the missing intensity values based on the values of existing neighbors in each of red blue and green channel.
Fig. 1.2 Bayer CFA pattern (left) & RGB intensity values from imaging sensor (right) [2]
Additional processing such as white balance, gamma correction, and contrast enhancement is performed before the eventual storage. Finally, the image is stored in the memory device in a particular image format (RAW, TIFF or JPEG)), but JPEG is the most common choice. A video is simply a sequence of images called video frames. So the steps performed in video acquisition are same as image acquisition but the resulting video frames are encoded using MPEG-x , H.26x or 3GP codecs.
Each step during multimedia acquisition introduces
various artifacts in the captured image. These artifacts are used in forgery detection and source camera identification.
References
[1] 1. J. A. Redi, W. Taktak and et. al., “Digital image forensics: a booklet for beginners”, Multimedia Tools and Applications, Vol. 51(1), pp. 133-162, January 2011.
[2] 1. http://www.i-cubeinc.com/3D/3shotcm.shtml