Goal: Histogram processing for image enhancement

• • Know how to enhance image by histogram processing

  • Apply look-up table to enhance images

  • Learn histogram equalization, histogram backpropagation, and mean shift

  • Practice image retrieval by histogram matching.

Readings and sample codes

• • OpenCV 4 Computer Vision Application Programming Cookbook, by D. M. Escrivá, R. Laganiere, Fourth Edition, Packt Publishing, 2019.    [Book URL]  [GitHub]

    • Chapter 4 Counting Pixels with Histograms (PDF, code)

  • See the README.txt in code for more details of sample codes.

  • You can use the provided sample codes in your homework, but you have to replace the images with yours.

Web Report 

• • Create a web page with description texts and a lot of pictures for the following programs

    • histogram.cpp: (1) Computing the image histogram. (2) Equalize the image. (3) Applying Look-up Tables to Modify Image Appearance. (p.104, p.105, p.110, p.112, p.115 in PDF) 

      • Change the image and the look-up tables to practice this sample code.

    • contentfinder.cpp: Backprojecting a Histogram to Detect Specific Image Content. (p.117, p.119, p.123 in PDF). 

      • Change images with blue sky or green grass to get your results.

    • finder.cpp: Using the Mean-shift Algorithm to Find an Object. (p.124, p.126 in PDF). 

      • Change the image with persons and use human's face as the target to apply mean-shot algorithm.

    • retrieve.cpp: Retrieving Similar Images using Histogram Comparison (p.130 in PDF) 

      • Collect a set of images with "similar colors". Use one image as query image, all the other images as gallary images. Retrieve similar image of the query from the gallery.

    • integral.cpp: Create a binary image by adaptively setting a threshold. (p.138 in PDF)

    • tracking.cpp: Locate an object in the image. (p.142 in PDF)

  • You have to give explanations of the codes. And you have to identify how did you modify the source code to get your results.

  • Submit your web address by Microsoft Teams.