DeepEarthMY Dataset

The DeepEarthMY dataset is a high-resolution remote sensing dataset specifically created for tropical land-cover semantic segmentation in Malaysia. It consists of 4007 annotated PNG images, each with a resolution of 1024 x 1024 pixels and an average size of 828 KB. The dataset captures diverse land-cover types, including roads, buildings, water bodies, forests, agriculture, barren land, and background. It is organized into two distinct environments—Urban and Rural—to represent different land-use patterns. Each image is manually annotated to provide high-quality labels, making this dataset ideal for training and evaluating deep learning models for semantic segmentation in tropical regions.