Object Centric vs Scene Centric Dataset

Hi Memebers,

Could you please help me to understand the difference between Object Centric vs Scene Centric dataset? Please correct me if I am wrong at some point listed in the following:

MNIST and Fashion-MNIST datasets are Object-Centric datasets because only 1 object that try to classify correctly whereas the Scene datasets are" Places397, MIT-Indoor67, etc. because there are multiple object in one image? Please correct me if i am wrong…

Is ImageNet dataset scene-centric or Object-Centric dataset? And what about Event8 scene dataset?

Thank you!