PASCAL VOC 2007 The dataset is suitable for image classification and object detection.

The dataset includes 9,963 images with 24,640 annotated objects. 50% data can be used for training or validation, and 50% data can be used for testing. The number of images and object categories in the training/validation set and the test set is roughly the same.

The labeled images in the training set are divided into 20 categories:

• 【人】：people
• 【Animals】: Bird, cat, cow, dog, horse, sheep
• 【Transportation】: airplane, bicycle, boat, bus, car, motorcycle, train
• 【Indoor】: bottles, chairs, dining table, potted plants, sofa, TV/monitor