TLDR
This post is about creating your own custom dataset for Image Segmentation/Object Detection. It provides an end-to-end perspective on what goes on in a real-world image detection/segmentation project. I created a dataset (currently open-sourced on Kaggle) which contains 81 image segmentations each for the five Simpson’s main characters (Homer, Lisa, Bart, Marge, and Maggie) It took me a day and a half from start to finish with just around 400 images.via the TL;DR App
no story
Written by mlwhiz | Data Scientist @Facebook.
Data science communicator at mlwhiz and TDS. Connect on Twitter @mlwhiz
