Description:

This dataset is designed for semantic segmentation in computer vision, specifically targeting the segmentation of swan objects. It provides labeled images for training models to distinguish between different regions within the images.

Download Dataset

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.

This dataset is sourced from Kaggle.