SCE Carleton Logo
  Carleton > Engineering > SCE > Faculty > A. Adler > Courses > BIOM5200 > Final 2007


BIOM5200 - Final Exam Winter 2007

You have 80 miutes to complete this exam. The exam has three (3) questions; you are required to answer any two (2) of them. Each question is worth equal marks. This is a closed book exam; however, you are permitted to bring two (2) 8.5"×11" sheet of notes into the exam. You are permitted to use a non-programmable calculator. You may not communicate with anyone during the exam except the instructor.

1. Ultrasound imaging and image restoration

    1A. An ultrasound system is used to image a baby in a mother's abdomen. A mechanically rotated transducer operating at 3 MHz is used to generate mode-B images. Sketch the setup of this transducer system, showing the transducer and its associated electronics. Describe how an image is formed.

    1B. A grandmother is studying the image and wants to know whether the baby has curly hair. Assume that baby hair has a diameter of 50µm. Define speckle noise. Will speckle noise interfere with the visibility of hair in this image? Explain.

    1C. Medical imaging systems have many sources of error, and it is the goal of image restoration to attempt to correct for these errors. Discuss the difference between heuristic and model based approaches to image restoration. Most medical imaging systems are better candidates for model based image restoration, than (for example) pictures of your vacation with a digital camera. Discuss this statement, and suggest reasons why this would be the case.

2. X-ray sources

    2A. An X-ray source generates a large number of low energy photons. A source typically includes a shield to remove these photons i) Discuss why these low energy photons have no diagnostic value and, ii) describe a technique by which they are eliminated from the X-ray beam

    2B. If the X-ray source is not a point source, the source distribution will introduce blur into the image. i) Sketch a diagram of the X-ray image formation in this case, ii) indicate the magnification of the blur, and iii) discuss how the magnification of the blur varies accross the target for thick targets.

    2C. Given a X-ray system used on thin targers, you are asked whether it is possible to use a Wiener filter to compensate for the blur due to the source size. i) Is the Wiener filter a good idea in this case, and ii) Discuss how you would set up Wiener filter, including any measurements you would make.

3. Contrast Enhancement

    Consider the 4×5 image represented in matrix form below. The region of interest (ROI) comprizes the centre pixels (with values 15,16,17). The surrounding region is the background.
    10 10 10 10 10
    10 15 16 17 10
    10 15 16 17 10
    10 10 10 10 10

    3A. i) Apply windowing to translate the pixels from m=5 to M=20 onto the range [0 .. 255]. ii) Calculate the average contrast ratio between the ROI and background before and after windowing

    3B. Apply histogram equalization to the image onto the range 0 to 255.

    3C. Apply a Sobel filter to detect horizontal edges in the image (only calculate the first two image rows). Comment briefly on the result. Assume the background region is replicated in the area surrounding the image

Last Updated: $Date: 2007-11-24 01:39:40 -0500 (Sat, 24 Nov 2007) $