OpenCV

Almost every tutorial here comes from PyImageSearch, Adrian Rosebrock's site, with a few from LearnOpenCV, Satya Mallick's site. Each post is a self-contained project with downloadable code. Good first projects use classic image processing and need no neural networks: Build A Document Scanner, Build A Barcode Scanner, Text Skew Correction, Automatic Target Detection and Object Tracking with Camshift.

Faces are a big theme. Work through Facial Landmark Detection, the faster 5-point landmark detector, Build A Face Detector using OpenCV and Deep Learning, Build a Face Recognition System, Learn Face Clustering with Python, EigenFaces using OpenCV and the Face mask detector, roughly in that order. Tracking projects include Tracking Multiple Objects with OpenCV, Dlib Correlation Object Tracking (2 parts) and People Counter, which combines detection and tracking to count people crossing a line.

The deep learning projects use OpenCV's dnn module to run pretrained models: Semantic Segmentation, Instance Segmentation, Object Detection using Mask R-CNN, Neural Style Transfer, Hand Keypoint Detection, and text detection with the EAST detector followed by OpenCV OCR with Tesseract. Image Stitching builds panoramas. Most posts date from 2017 and 2018, so pin your OpenCV version if a function signature has changed.