Scan - Print Do repeated scans and prints to show image degradation. HW0202.

Semester Work Chart

Element Short Description HW0202 –1 Scan-Print Sequences PJ1A Scan-Print Improvement by Adjustment PJ1B Color Printer Improvement by Adjustment PJ1C Explain LUT HW0209A –2 Bone Marrow Filtering and Image Quality HW0209B –2 JPEG and Tiff compression and Image

Quality HW0216A –3 Digital Camera Image Improvement HW0216B –3 Contrast, Brightness, Equalization

Algorithms PJ2A Recap on Figure 3-25 PJ2B Rolling Ball and Top Hat HW0223A –4 Three Sines and FFT HW0223B –4 Muscle Striations and FFT PJ3A Scion Macro PJ3B Glare removal PJ3C Matlab Isometric Plot from BMP image HW0301 –5 Rice grains PJ4A Locate randomly oriented rice grains PJ4B Three filters/processors HW0308 –6 Segment retina.tif HW0322 –7 Feature measurement PJ5A Villi ‘grains’ evaluation HW0329 –8 Automated Bug Inspector HW Optional* Stereology PJ5B To Be Determined – Make a Proposal PJ6B WiT HW0419 –9 Stereo brain orientation HW0426 –10 To Be Determined *May be used for extra credit on HW.

Scan - Print

• Do repeated scans and prints to show image degradation.

HW0202

Homework for 2/9

• Locate the “bone marrow” image shown in Figures 2-5 of your textbook.

• Acquire a background image and subtract it from the original or divide it into the image. Comment on what you observe. Which is preferred -- divide or subtract?

• Perform spatial averaging over 33, 55, and 1111 neighborhoods, and plot the standard deviation of pixel values. Find s.d. from UTHSCA.

HW0209A

Suggested Readings from Jasc

• Before You Print

Suggested Readings -- Jasc

• Sample Jasc Palette

• About the Jasc Palette File Format

• Setting Palette Transparency

Suggested Readings

• Using the Hue/Saturation/Lightness Dialog Box --Jasc

Colour_interchange.pdf.lnk

Reading AssignmentRuss, Chapter 2

In Class Mini-assignment

• Try different levels of JPEG and TIFF compression on a busy and a ‘smooth’ image.

• Do image blowup and describe differences.

• Do subtraction from original and describe differences.

HW0209B

In Class Mini-assignment

• Compare 9:08 and 9:10– Do a subtraction– Also improve image quality– Use histogram equalization

• what does it do?

HW0209B

Assignment for next week

• Read Chapter 3

• Finish in-class assignment

Bone Marrow Imagesfor 2/9

• Bone Marrow and Bone Marrow2 (BoneMarr, BoneMar2): Light microscope images. Use local equalization to increase contrast between dark features in the original Bone Marrow, compare this video image to the higher resolution image obtained with a digital camera (BoneMarrow2). Compare smoothing and median filters on noise and feature shape.

HW0209A

Read next chapter in Russ(Chapter 5)

• For 2/14/00

Homework#1 for 2/16/00

• Pick any one of the digital camera photos provided, and use any image enhancement techniques to produce a best quality reproduction of the original subject. The end result should be observable on a gamma corrected monitor. Use any of the filtering and enhancement techniques described in the course or others that may be available.

• Turn in a softcopy of the processed image and a description of the corrections applied and their sequencing.

HW0216A

Homework#2 for 2/16/00

• Given the following image 5 level image

• Design, apply, describe and defend routines for contrast and brightness adjustment, and for histogram equalization routine. Apply to this image with a 5 level result.

1 2 2 1 11 3 4 1 11 2 2 1 11 1 0 1 11 1 1 1 1

HW0216B

HW#2 Suggestions(for 2/16/00)

• Construct an image histogram/distribution• Define the target desired histogram/distribution

– may be discussed in class

• Select a level reassignment algorithm to approximate the target.– For this assignment, each level must be reassigned to

only one level in the result.

• Construct an actual image histogram/distribution• Construct a new image array

HW0216B

HW#1 for 2/23• Start with threesins image which is 3.31” square

– consists of 3 sinewave images added together

• Prepare a circular filter .514” diameterr to eliminate the two lower fequencies

• Repeat but eliminate high frequency by an inverse of the previous mask.

• Provide copies of the original and two resulting images.

HW0223A

HW#2 for 2/23

• Consider muscle2.tif• Use a mask or combination of masks with the

FFT to emphasize periodicity of the structure in the resulting image.

• Suggest additional processing to achieve best result

• Provide a copy of the before and after image and describe the mask(s) and processing.

HW0223B

HW#2 continued

• Determine the spatial frequency of the muscle striations.

• Give the units of measurement.

• Explain how determined.

HW0223B

Correlation

• Pick a rice grain. Use correlation to locate similarly oriented rice grains in rice.tif.

HW0301

Segmentation for 0308

• Segment retina.tiff to find the vessels. Describe fully the process used for segmenting.

Features HW0322• Begin with the image that shows the region

along the villi of mouse intestine.

• Process as shown in Fig 10, p439 with added cleanup for measurement purposes.

• Then do an IT analysis on the villi that fall entirely within the image.

• Provide a statistical summary of means and sd’s for the villi properties as determined from IT

Optional

• For each of the following, give the stereological notation

• Describe any added conditions for accurate measurement of the notated variable.

• Identify what densities can be determined with the aid of the notated variable.

• Continued

Optional

• 1. Total area fraction of segmented black pixels.

• 2. Formfactor: 4*pi*Area/Perimeter^2

• 3. Consider how to use Image Tool to calculate, directly or indirectly, the sample densities described earlier: NV, AA, IS, SV

The Automated Bug Inspector

• A proposal has been made to develop an automated image analysis system for instecting bugs. You are asked to develop such a system. The initial assumptions for development are:

• 1. This is always one and only one bug in each image.

• 2. The bugs are always aligned in a standard position.

• 3. The bug antennae are thinner than bug legs.

HW0329

To determine

• When feasible, the system is to be broadened to analysis of images for which these assumptions are not valid. Measurements to determine for each bug are:– Overall bug size

– Number of legs

– Length of each leg

– Number of antennae

– Length of each antenna

HW0329

Approach

• Consider pre-processing, segmentation, and analysis.

HW0329