A lightweight dataset of dandelion flowers for simple machine learning.
The dataset was created by cropping flowers of Dandelions (Taraxacum officinale L.) from downward-facing images collected at about 1 meter height. Each tile is 128 by 128 pixels, which is about the size to fit an average flower for an image collected at this distance from the ground using a smartphone with 3264 × 2448 image resolution.
The dataset is aimed at educators that typically need for a straight forward and lightweight dataset. The dataset has limited application since all images were collected on the same field on the same day, but it should be able to do a decent job in common scenarios. I'm pretty sure that most of you are familiar with the following scenario:
Image collected by Andres Patrignani on 18-April-2021 in Manhattan, KS.