Skip to content

Latest commit

 

History

History
25 lines (18 loc) · 1.4 KB

File metadata and controls

25 lines (18 loc) · 1.4 KB

Keras UNet implementations (Github)

Semantic segmentation

Semantic segmentation

The Jaccard coefficient measures similarity between finite sample sets, and is defined as the size of the intersection divided by the size of the union of the sample sets. Ground truth vs. predicted bounding boxes on a stop sign
Areas of overlap and union
Comparison of results

Getting started

Creating a virtual environment

pipenv install

About the dataset

For image segmentation we need images to work on and masks that provide the ground truth for segments in the images. A mask is an area in the image belonging to a specific class. In this dataset they are provided in the train_wkt_v4.csv file that has a MultipolygonWKT column which contains MultiPolygon objects of the Shapely library.