Data Format: Semantic Segmentation

Introduction

Semantic segmentation is an image problem that identifies the content and location within images, typically requiring identification of all pixels of objects in the image. When defining the data format, each image corresponds to one image annotation in a one-to-one relationship, where pixels in the annotation represent labels for the corresponding pixels in the image.

In the annotation images, the background is labeled as 0, and each type of annotation is labeled as 1, 2, 3... and so on.

Example

image_Label,image_Source
label/001.jpg,images/001.jpg
label/002.jpg,images/002.jpg

FAQ

1. The source image and annotation image must have exactly the same dimensions and orientation