My path towards learning and using AI 

Formally trained as an astrophycisist, it was challenging for me to begin to work on everything concerning AI. As a researcher, it has been challenging to use current tools as black boxes, so here are my little "deeper dives" into everything concerning AI.

Labeling 

One point in common between observational astrophysics and AI is the capability we should have to handle big data sets. It is not surprising then to acknowledge that I felt comfortable gathering images and labeling them, in a project that was intended to carry out object detection. But as I knew absolutely nothing about this matter, I had to learn a bunch. These are the sources that helped me better understand how to label right, and the importance beyond a proper and consistent labeling.

Stop these mistakes in image labeling for object detection (labellerr.com)

The Effect of Improving Annotation Quality on Object Detection Datasets: A Preliminary Study (thecvf.com)

[1811.00982] The Open Images Dataset V4: Unified image classification, object detection, and visual relationship detection at scale (arxiv.org)

Differences between image classification and object detection:

A Gentle Introduction to Object Recognition With Deep Learning - MachineLearningMastery.com