Image filtration

Try to improve on your tubule count from your previous assignment by adding following filtration operations or their combinations into your processing. Discuss the results:

• blur filter

• median filter

• edge detection filters

• bilateral filter

What filters or combinations improved on the results? Discuss, which of the filters are convolutional filters.

Histology image correction

Dataset

Dataset link: link (Histology nuclei)

H&E and P63 registered images: H&E , P63

The stained colors of the tissue components are popularly used as features for image analysis. Dataset contain stained histology images. Variations in the staining condition of the histology slides could however negatively impact the accuracy of the analysis. 

The assignment goal is to experiment with the histology images and try to correct the color distribution of the provided dataset.

Experiment with the following tasks for image correction:

• Color space conversion - Grayscale, RGB, YCbCr, HSV, XYZ, Lab

  • Histogram computation (visualize histogram for each color model used)

  • Observe each color channel ranges - Are any color spaces different? 

• Histogram normalization - for multiple color models, experiment with normalizing the histogram values. Observe the results.

• Source -> Target color correction using eCDF (effective Cummulative Distribution Function) and linear interpolation

  • Set target as an image you selected from the dataset. We want to change the color distribution of other (sources) images to the target one

  • Experiment for Grayscale, RGB and YCbCr color models only

Tips & Tricks

bilateralFilter()

Operations using histogram:

calcHist()

normalize()  (histogram)

compareHist()

Contrast correction:

pow()

Color space conversion:

cvtColor()