Oceanic Whitecaps Segmentation Data

The dataset comprises sea surface images captured by the Shanghai Ocean University's R/V Songhang in the open sea of the Northwest Pacific. Oceanic whitecaps were segmented through a combined threshold-based and manual annotation approach to generate deep learning training samples. Within the dataset: The JPEGImages folder contains sub-images extracted from the original photographs. The SegmentationClass folder stores corresponding pixel-level whitecap annotations, where whitecaps are labeled as 1 and background as 0.