ayhokuyan / CS484-555-Introduction-to-Computer-Vision

Homework Assignments for CS484-555 Introduction to Computer Vision, Bilkent University

CS484-585 Introduction to Computer Vision

The assignment implementations for CS555 Introduction to Computer Vision course, Bilkent University.

Coverage of the repository

HW1

Image Operations and Morphological Operators

• Compute and show histograms of grayscale images.
• Apply histogram equalization to enhance contrast.
• Apply manual and Otsu's thresholding and observe the differences.
• Implement morphological operations such as dilation, erosion, and further, opening and closing. Apply these techniques on example images to determine the number of elements using Connected Component Analysis (CCA).
• Implemented in Python (Jupyter Notebook)

HW2

Neural Networks

• Train a neural network from ground up. completing the given template.
• Use the functions implemented to for a Cat vs Non-Cat binary classification problem.
• Implemented in Python (Jupyter Notebook)

HW3

Edge Detection and Hough Transform

• Obtain horizontal and vertical gradients using Sobel and Prewitt operators, apple edge detection by finding the magnitude of gradients.
• Apply Canny Edge Detector with the steps,
  • Gaussian Smoothing
  • Edge Filtering
  • Norm and Gradient Calculation
  • Non-maximal Suppression
  • Hysteresis Thresholding
• Use scikit-image implementation and compare the results.
• Report best out of various combinations of parameters.
• Implement and use Hough Transform to find straight lines in pictures, show the images and the corresponding Hough space heatmaps. Used both the theoretical algorithm and the simplified version to decrease the time complexity.

Project:

For project, see CartooNet