Fundamentals of image smoothing

Install Anaconda to create a virtual Python environment and run a Jupiter notebook; verify Python 3.7, launch conda, and test a simple print in notebook.

Learn to install OpenCV on macOS by setting up Python and Anaconda, installing Xcode and command line tools, and using Homebrew to compile and verify OpenCV for image manipulation.

Install OpenCV part II: verify dependencies, install in a conda environment, and run in a Jupyter notebook to import the OpenCV package and check the version to confirm it works.

Learn the basics of image storage and simple operations like increasing intensity and reducing density to build a solid foundation before diving into convolutions.

Explore basic image processing with OpenCV by loading images, analyzing color versus grayscale representations, and understanding images as pixel arrays with RGB channels and 0–255 intensities.

Learn how to inspect image properties and manipulate regions using a 3-channel BGR image, including splitting into four quadrants, displaying them, and reassembling while handling the top-left coordinate origin.

Explore image math by adjusting pixel densities with add and clip operations, brightening or darkening images within 0–255, using matrix-based operations and display results.

Convolve a 3x3 kernel with an image by sliding, multiplying, and summing. Pad the image to preserve its 5x5 size, maintaining output dimensions.

Walks through the convolution code for a grayscale image using a kernel, padding to keep the output size, border replication, and region of interest with element-wise multiplication and sum.

Apply convolution kernels to color and grayscale images, including blur, emboss, outline, Sobel, and sharpening, using 3x3 and larger kernels to emphasize edges and smooth details, demonstrated in a notebook.

Apply average smoothing using 3x3, 5x5, or 7x7 kernels to perform a convolution that averages surrounding pixels, reducing noise and smoothing details in images.

Explore gaussian blur as a weighted smoothing technique where weights fall off with distance from the center using a sigma-based distribution, offering edge-preserving smoothing compared with simple averaging.

Apply median blur to smooth images by taking the median of the neighborhood, ignoring outliers and removing salt-and-pepper noise with configurable square kernels (3, 5, 9), in grayscale and color.

Explore bilateral blur, a filter that preserves edges while smoothing noise by weighting neighborhood pixels with color and spatial distributions, using diameter and color/space standard deviations.

Practice applying smoothing techniques, especially bilateral filters, to diverse images; adjust kernel sizes from 3x3 to 5x5 and explore standard deviation; implement with OpenCV and Python for batch image processing.