Labels in the Dataset:
- Background (Label “0”): Represents the areas of the image that do not contain any swan objects.
- Swan (Label “1”): Represents the regions of the image that contain swan objects.
The dataset includes a variety of images featuring swans in different environments and poses, ensuring comprehensive coverage for effective model training. The high-quality annotations enable precise segmentation of swan objects from the background, making it a valuable resource for developing and evaluating semantic segmentation algorithms.