Update README.md

README.md

@@ -5,6 +5,20 @@ size_categories:
 - 1K<n<10K
 ---
 # Danbooru SFW 512 Crop Dataset V1
+This dataset includes image and bounding box pairs where the images are samples from 
+Danbooru 2021 SFW and the bounding boxes are crops of the central character(s). It can 
+be used to train a model that automatically crops the base dataset in order to create a cleaner
+and more character-centric version. Importantly, the crops aim to remove artifacts such as
+horizontal/vertical bars that could negatively impact performance on image generation tasks.
+
+## Bounding Boxes
+The base images are all 512x512, the bounding boxes represent crops with a maximium size of
+512x512 and a minimum size of 128x128. Each bounding box is represented as a 3-dimensional 
+vector [x, y, l] where x is the left coordinate of the box, y is the top coordinate of the box,
+and l is the side length of the box. These elements have been converted to floats and 
+normalized according to their maximum possible values (512). For example, [16, 28, 256] would 
+have been converted to [16/512, 28/512, 256/512], or [0.03125, 0.0546875, 0.5]. Using this 
+conversion, the minimum value of x, y, and l is 128/512, or 0.25.
 
 ## Original Dataset Citation
 ```bibtex