UNIVERSITY OF COLOMBO, SRI LANKA

UNIVERSITY OF COLOMBO SCHOOL OF COMPUTING

BACHELOR OF COMPUTER SCIENCE and BACHELOR OF INFORMATION & COMMUNICATION TECHNOLOGY Third Year Examination in Computer Science (2005/ 2006) Semester I - 2009 SCS3003/ ICT3009 Computer Graphics and Image Processing

(2 HOURS)
Answer ALL questions. Non-programmable calculators are allowed.

No of Pages = 05 1.

No of Questions = 04

(a). Briefly describe 3 recent application areas of Computer Graphics explaining the techniques they use. (6 marks) (b). What is a scan line? In edge walking polygon rasterization, the algorithm moves across a scan line and performs a quick check to determine if each pixel is inside or outside the polygon. Explain this algorithm considering the scan lines A and B given in the figure below.

A B

(6 marks) (c). What is Window to viewport transformation? Suppose you have a window with coordinates (wxmin, wxmax,wymin, wymax) and a viewport with coordinates (vxmin, vxmax, vymin, vymax). Derive a transformation that converts a point (x,y) in window coordinates to a point (x,y) in viewport coordinates. (5 marks) 1

(d) Suppose we wish to clip lines to a rectangular window with a rectangular hole in it:

The shaded region is our window

A set of lines before clipping

A set of lines after clipping

Call the outer rectangle R with vertices r0, r1, r2, and r3, and call the inner rectangle H with vertices h0, h1, h2, and h3. You may assume that hi is the corner closest to ri. Describe an algorithm for clipping lines to this object. (8 marks)

2. (a). Why do we use a 4x4 homogenous coordinate transformation matrix when describing 3 dimensional geometric transformations.? (3 marks)

(b). Provide a 3x3 matrix that will compute the new vertices of a planar house centered at [10, 5] after a rotation of 90 degrees about its center:

(6 marks)

(c). Compare advantages and disadvantages between ray tracing and polygon based rendering. (4 marks)

(d). The shading intensity at any given point on a surface is, in general, comprised of three contributions, each of which corresponds to a distinct physical phenomenon. List all three, stating how they are computed in terms of the appropriate vectors. (6 marks)

(e). Both Gouraud and Phong Shading interpolate along polygon edges to compute intensities. Describe the difference between Gouraud shading and Phong shading, in terms of interpolation? (6 marks)

3. (a) Compare the Canny edge detector and the Laplacian-of-Gaussian (LoG) edge detector for each of the following questions. (Please answer very BRIEFLY.) (i) Which of these operators is/are isotropic and which is/are non-isotropic (i.e., directional)? (2 marks) (ii) Describe each operator in terms of the order of the derivative that it computes. (2 marks) (iii) What parameters must be defined by the user for each operator? (2 marks) (iv) Which detector is more likely to produce long, thin contours? Briefly explain. (3 marks) (v) Give two major reasons why the Canny edge operator is generally preferred over the Laplacian-of-Gaussian (LoG) edge operator for most applications. (4 marks) (vi) Describe a major advantage or disadvantage of using an isotropic operator instead of a non-isotropic operator with respect to the Computational efficiency and noise in the image. (4 marks) (vii) What is the purpose of non-maxima suppression and hysteresis that are done in the Canny edge detector? (4 marks)

(b) The two texture images shown below are of size 80x80, with black (0) and white (1) pixels. What can you say about their intensity histograms? Explain your answer.

(4 marks)

4. (a) Median filtering gives better results than neighbourhood averaging in image smoothing. Explain the reasons for this using an example. (4 marks) (b). When a set of images is examined, the following problems are found: (i) bright, isolated dots that are of no interest (ii) lack of sharpness (iii) not enough contrast in some images Describe the image enhancement techniques that can be used to correct these problems. (6 marks)

(c). Briefly explain the method for detection of straight lines using Hough Transform. (5 marks)

(d) A certain inspection application gathers black & white images of parts as they travel along a conveyor belt. It is necessary to sort the parts into two categories: parts with holes and parts without holes. An example of an image that might be taken by the inspection camera is shown below. Propose a method to identify and locate the objects of each category in the image so that they can be picked up by a robotic system and placed in different bins. Assume that the imaging system knows where each image pixel is located on the conveyor belt at every point in time. Provide a flow chart of the algorithm you propose.

(10 marks)