jomoll/silent-heart-dataset-p003

该数据集包含图像和对应的描述，图像可能包含心脏（heart）或非目标物体（decoy）。每个图像都有标签指示是否为心脏，以及其他相关信息，如背景类型、心率、左心室率、物体形状、命名物体、静音物体、特殊物体等。数据集分为训练集、验证集和测试集，分别包含400,000、4,000和4,000个示例。

The dataset consists of images and their corresponding captions, which may contain a heart or non-target objects (decoy). Each image is labeled indicating whether it is a heart, along with other information such as background type, heart rate, left ventricular rate, object shape, named objects, silent objects, special objects, etc. The dataset is split into training, validation, and test sets, containing 400,000, 4,000, and 4,000 examples respectively.