"Writ in Water": Passionate Entrance for Tears and Impassive Exit on Boat

This synthetic dataset comprises computer vision training data designed for multi-task learning applications including object detection, semantic segmentation, and depth estimation. The dataset features scenarios involving water-based environments with emphasis on dynamic human interactions, specifically entrance and exit sequences in aquatic settings captured from boat-mounted perspectives. Generated using synthetic data generation techniques, it provides pixel-perfect ground truth annotations across multiple vision tasks, enabling robust model training without privacy concerns or real-world data collection constraints. The dataset is intended for researchers developing AI systems for marine surveillance, autonomous watercraft navigation, and emotion-aware computer vision applications in challenging water-reflected lighting conditions.