=========================preview======================
(ELEC360)final02f.pdf
Back to ELEC360 Login to download
======================================================
Hong Kong University of Science & Technology Department of Electrical & Electronic Engineering
ELEC 360: Digital Media & Multimedia Applications
Fall 2002 Lecturer: Chow King Wai
Final Exam (December 12, 2002)
Name: _________________________________________
Student ID: _____________________________________
E-mail: _________________________________________
Signature: ______________________________________
Closed book & closed notes. Calculator is allowed provided that it does NOT
store and/or output text data.
When you are in doubt, please raise your hand and ask as soon as possible.
Digital camera is a prime example of digital multimedia system which touches on a good number of multimedia elements and makes use of a variety of associated techniques. Even though this topic was already covered in a lecture a couple of weeks ago, it is worthwhile to spend some more time examining each feature in greater details.
The primary objective of a digital camera is to take still pictures and save them in a digital form. The sensor plays a crucial role in the system as it turns photons into electrons. After the A/D stage, each spatial sampling point or pixel is represented by a digital number which corresponds to some light intensity level. In general, each raw pixel value can only represent one of the three primary colors (red, green and blue). The resulting spatial structure is called Bayer pattern as shown in Figure 1.
Figure 1: Bayer Pattern
1.
Why is it that the raw pixel *cannot* take on all 3 primary color values?
2.
In order to provide a complete RGB representation for each pixel location, color recovery or interpolation is used to estimate the `missing color component values for each pixel. Consider the pixel region in the shaded area of the following Bayer pattern, apply linear interpolation using all the nearest neighbors of the same component type and compute the `missing color component values.
3.
After examining the result obtained in Question 2, we discover that the method used above has certain drawback. What is it? Propose additional procedure or steps which reduce the impact of the said drawback. Feel free to use functional block diagram to illustrate your approach.
4.
Even though the direct output (voltage level before A/D or digital value) from the imaging sensor corresponding in a linear fashion to the amount of incident light
G 10 B 25 G 30 B 200 G 33 B 199 G 31
R 200 G 20 R 210 G 20 R 30 G 25 R 29
G 12 B 27 G 25 B 230 G 27 B 222 G 20
R 210 G 19 R 220 G 23 R 33 G 31 R 35
G 15 B 19 G 28 B 210 G 20 B 250 G 19
R 195 G 21 R 222 G 22 R 29 G 10 R 33
G 18 B 30 G 24 B 200 G 22 B 215 G 25
(i.e. amount of photons), it is necessary to perform gamma correction. The reasons are two-fold. Human visual system perception operates in a non-linear scale. Gamma correction transforms the raw image data into a perceptually uniform domai