COCO is a widely used dataset for object detection , image classification, and semantic segmentation tasks. It is a benchmark dataset which is used in different academic and industrial research . Although it is so useful many people struggle to use this dataset because the huge size of the whole COCO dataset makes it difficult to download as It uses up a lot of our internet data along with unnecessary space in our system. Whereas we might just be interested in using just a subset instead of a whole dataset. This blog here is to present how to download a customised data with specific annotations.
This post is written by one of our Data Scientist experts Mahima Modi on Medium medium.com/@mjmodi19
Original Source: Calibrating Coco dataset for object detection