make-sense

Why we included this project

Building a computer vision dataset usually means hours of drawing bounding boxes by hand, and makesense.ai exists to take the worst of that work away. It runs entirely in the browser, so there is no installation and no dependence on your operating system; you upload a folder of images, mark up the objects or people in them, and export the annotations in one of the formats your training pipeline expects. The built-in AI can do some of the drawing for you: a pretrained SSD or YOLOv5 model will pre-draw boxes and suggest labels, and a PoseNet integration estimates body joints for pose work. For small to medium projects it is a practical alternative to paid labeling platforms, giving you fast annotation without any infrastructure to stand up.