Blog

Learn how to generate and display HDR images from an exposure sequence in OpenCV. We cover Debevec, Robertson, and Mertens exposure fusion algorithms with camera response function estimation.

Learn how to remove small noises, strokes, and damage from old photographs using OpenCV's cv.inpaint(). We cover the Telea and Navier-Stokes inpainting algorithms.

Learn about Non-local Means Denoising algorithm to remove noise from images using OpenCV. We cover cv.fastNlMeansDenoising(), cv.fastNlMeansDenoisingColored(), and multi-frame denoising.

Learn to use cv.kmeans() function in OpenCV for data clustering. We cover single feature, multi-feature data and color quantization of images using K-Means.

Learn the concepts of K-Means Clustering algorithm. We go through the step-by-step algorithm using a T-shirt size problem to explain how iterative centroid updates work.

Learn to use SVM with HOG (Histogram of Oriented Gradients) features for OCR of handwritten digits. This approach achieves ~94% accuracy, better than kNN with raw pixels.

Learn an intuitive understanding of Support Vector Machines (SVM). We cover decision boundaries, support vectors, margin maximization, and kernel trick for non-linearly separable data.

Learn to build a basic OCR application using kNN knowledge. We use OpenCV's digits.png dataset to recognize handwritten digits achieving ~91% accuracy.

Learn the concepts of the k-Nearest Neighbour (kNN) algorithm. We cover classification, feature space, weighted kNN, and implement a simple example with OpenCV.