Python Digital Image Processing From Ground Up™

Download python from python.org and install the windows 64-bit executable, choosing the appropriate 3.x version and the x86_64 installer for your system. Remember that 32-bit versions have memory limits.

Install Python 3 7 by following the on-screen steps, add Python 3 7 to path, and set the install folder to documents/python 3 7. Verify with print Hello Python.

Learn to use Idle, Python's interactive editor, to write, save, and run simple scripts with print statements and quick tests, then create and manage projects with import and export.

Install matplotlib, numpy, and scipy with pip in the python 3.7 script directory, using the command prompt, to enable plotting and numerical processing for this dsp course.

Open IDLE, create and save a Python file in the DSP course folder, and use Python 3's print function with parentheses while learning indentation, automatic variable assignment, and if statements.

Explore Python variable types and the object-oriented, dynamically typed nature of the language, declaring integers and floats, casting to float, manipulating strings and concatenation, and printing results.

Learn how to declare and append to Python lists, access elements by index, and iterate over items with a for loop to print values.

Explore Python operators from basic arithmetic to exponentiation, modulo, and list concatenation, and learn to join strings with numbers using casting and printing.

Explore conditions and loops in Python by building if statements, using else branches, and applying the in operator with lists to control program flow.

Explore Python loops by demonstrating for loops and while loops, using range to iterate numbers and lists and print examples with step sizes.

Master while loops in python through hands-on examples: print 0 through 4, apply break and continue, and skip even numbers with a range-based for loop.

Master python functions by defining with def, using indentation to create blocks, and calling simple functions with and without arguments. Learn to print inside functions and return values.

Explore dictionaries as a data type with keys and values, and learn to create, initialize, access, iterate, print, and delete entries in Python.

Explore classes and objects in Python, learning how a class is a template that encapsulates variables and functions. Practice creating objects and accessing their members.

Explore applications of image processing across diagnostics, military object detection, consumer rendering, and biometric security. Understand how a digital image is a discrete 2D data grid of pixels with row and column indices m and n.

Explore image color from grayscale to rgb, and understand hue, saturation and intensity, plus spatial resolution, temporal resolution, bit resolution, and dynamic range.

Explore popular image formats like jpeg, gif, bmp, png, and tiff, and differentiate image data types—binary, grayscale, rgb, and floating point—to guide format choice in image processing.

Learn how to read, display, and convert images to greyscale using the Python Imaging Library (PIL), with hands-on file setup and code examples.

Convert a list of images from jpeg to png using the Python Imaging Library (PIL), by iterating over file paths, changing extensions, opening and saving with error handling.

Learn to perform basic image manipulations in Python using the PIL library, including opening an image, resizing to 150x150, cropping with a defined box, and rotating by 45 degrees.

Analyze an rgb image using numpy and matplotlib by loading it as a numpy array, inspecting its shape and dtype, and displaying with imshow and plt.show.

Learn to plot descriptive images by adding lines, dots, and asterisks to mark regions of interest, using plot functions, red symbols, and titles for clearer communication.

Add interactive annotation to an image display using the Python Imaging Library and pileup, letting users click four points and print their coordinates for region or area of interest.

Explore low-, mid-, and high-level image processing operations, including sharpening, noise removal, edge detection, binarization, contrast enhancement, and segmentation, and see how these techniques support analysis.

Perform binarization by applying a binary threshold to a greyscale scanned image, assigning 0 for pixels below 150 and 255 for others, using scikit-image data and matplotlib for visualization.

Explore image topology and key terms such as neighborhood, adjacency, path, and connectivity. Learn how 4-adjacent and 8-adjacent pixels form the neighborhood and connect via paths and components.

Describe machine vision systems for facial recognition, outlining acquisition, pre-processing, segmentation, feature extraction, and pattern classification to convert facial features into a feature vector for recognition.

Examine image histograms and histogram equalization in Python digital image processing, learning how to plot pixel value frequencies, analyze grayscale contrast, and distinguish foreground from background in an 8-bit image.

Demonstrates histogram equalization to improve image contrast by spreading a skewed histogram. Shows how the process uses the original histogram to transform the image into an equalized version.

Compute the histogram of an image using the python image library, convert to grayscale, flatten the array, and adjust the bin count.

This lesson demonstrates histogram equalization on a grayscale image using a Python workflow: load, convert to a 1d array, compute histogram and cdf, apply equalization, and compare results.

Explore how adaptive thresholding overcomes global thresholding limitations by computing local thresholds from image neighborhoods, using mean, median, or max-min statistics plus a constant offset.

Learn how geometric operations reposition image pixels to alter geometry, using rotate, flip, crop, and resize for distortion correction, effects, and image registration, guided by mapping functions and interpolation methods.

Explore how affine transformations map image coordinates using linear combinations and transformation matrices, covering translation, rotation, scaling, and shear with practical examples.

Explore image enhancement as processing that reveals task-relevant information by reducing noise and sharpening features. Learn spatial domain filtering, linear and nonlinear filters, and local neighborhood connectivity.

Learn how linear spatial filters use a kernel to weight neighborhood pixels, slide across the image, and sum the weighted values to form the new image.

Learn to perform gamma correction on images by loading with Pillow, converting to an array, normalizing by max, applying a gamma exponent, and scaling to 0-255 for display and save.

Explore grey level transformations to enhance images using linear and nonlinear mapping, where the pixel value r becomes s via s = c r + b, affecting contrast and brightness.

Perform gray-level transformations on a grayscale image in Python by applying negative, clamp, and square operations, then display results. Experiment with different constants and observe how pixel values change.

Perform pixel-by-pixel arithmetic and logical operations on images, including addition and subtraction, to detect differences, using additive or subtractive offsets and normalization or truncation to keep within 8-bit range.

Explore neighborhood processing in image analysis, applying linear or nonlinear operations to a reference pixel and its neighbors. Understand how outputs arise via neighborhood operations, convolution, correlation, and Marzook coefficients.

Explore how convolution and correlation process images by sliding weights over pixel neighborhoods and computing sums of products, demonstrated with a 1D array example A and B.

Explore 2-D convolution and correlation in image processing by flipping the 3x3 mask, applying it across the image, and observing effects like blurring, sharpening, and edge detection.

Learn how low-pass filters smooth images by attenuating high-frequency components in the spatial domain, using mean filters, convolution, and masks from 3x3 to 15x15.

Filter images in Python using the Python Imaging Library, applying filters like detail, edge enhancement, find edges, and custom kernels. Compare original and filtered results side by side.

Apply a 5×5 mean filter to a grayscale image using convolution, normalizing the kernel to average values, and generate a blurred mean profile image as the output.

Learn how to apply the minimum filter to a grayscale image using Sipi in Python, and compare its blurring effect with other filters for noise reduction.

Explore the performance of the maximum filter on a noisy image in python, compare it to the median filter, and learn why the maximum filter may not restore detail.

Apply a median filter to a grayscale noisy image to reduce salt-and-pepper noise, using a 5x5 footprint and reflect boundary handling; compare results with a mean filter.

Explore how mathematical operators enable edge detection in images by convolving masks like Prewett, Sobell, Keusch, and Robinson for vertical, horizontal, and directional edges.

Explore edge detection with the Prewitt mask by applying a vertical operator to a grayscale image, then compare it with a horizontal operator to reveal edges.

Apply the sobel edge-detection filter to a grayscale moon image using skimage, convolving the input with the sobel kernel and saving the edge-enhanced result.

Apply Sobel edge detection with OpenCV by loading an image, converting to grayscale, and computing Sobel x and Sobel y with a 5×5 kernel; view the results.

Learn to perform Laplacian edge detection with OpenCV by importing cv2, reading an image, and applying cv2.Laplacian with depth -1. Compare outputs with other image libraries.