Intelligent Systems Coursework Assignment

Intelligent Systems Coursework Assignment
Spring Semester, 2005
Write a report about working on the following tasks. The report is expected to have
following criteria:
Abstract of the problems and objectives
Clarity, structure and professionalism in the presentation of the report
Appropriate discussion and analysis of the problems in the tasks
Description of methods used and developed
Discussion on potential applications of the work carried out
Good use and understanding of appropriate references
Well argued discussion and conclusions
Screen shots of the results
Original codes developed should be attached to the report.
Task 1. (25%) Please implement algorithms in listing 5.14 (enlarge), 5.15 (shrink),
5.16 (xReflectionInPlace), 5.17 (average) and the subtraction described on page 100.
The final systems (can be individual systems or integrated) should have appropriate
interface to load input images and display the results.
Reference suggestion: sample codes in chapter 05 such as Dither can be used.
Task2. (10%)
Running the ConvolutionTool programme, try to use different types of kernels,
provide a discussion on the relationships between the kernel values and the effect on
output image (smooth or edge detection etc?)
Task 3.(40%) Please design and implement a programme that is able to compare the
distance (similarity) between two images based on their histograms. Try to use pairs
of images that either a) similar content, but different spatial arrangement; or b)
visually similar, but different colour (grey level) contrast. Please provide discussion in
the report, the analysis of the results and suggest what might be a better solution for a
more precise measure.
Reference: relevant lecture notes on histogram, content in chapter 6 and related codes.
Task 4. (25%) Please design and implement a segmentation algorithm, that will take
input image and produce a binary image. The centroid of the segmented objects in the
image is calculated. You can experiment with images which only contain one simple
Note: the coordinate of centroid is the average of all x coordinates and y coordinates
along the boundary of the object.