ORCA aims to advance marine visual understanding for machine learning models by formulating tasks that align with established computer vision objectives and by providing instance-level captions and bounding boxes to support fine-grained comprehension.
Supported Tasks
Object Detection: Identify and localize target objects within a predefined set of categories.
Open-Vocabulary Object Detection: Identify and localize target… See the full description on the dataset page: https://huggingface.co/datasets/WongYukKwan/ORCA.
