General Machine Learning Notes

The Ancient Secrets of Computer Vision 1 by Joseph Redmon

Level Types
Low Working on pixel values of images/videos, usually in the first stages of a CV pipelines: resizing, converting to grayscale, altering exposure, changing saturation, hue, edge detection, orientation, segmentation of important areas.
Mid Tie images to others and the real world: panorama stitching, created 3d models using light scans, find distance between objects, object motion detection, create time lapses.
High Tie in previous level techniques to bring semantic information to the process: image classification, object detection, semantic segmentation, instance segmentation.