Does this DoTA dataset contain optical flow features? Does it contain the features of the vehicle's ego-motion?
Do we need to preprocess the data set to obtain the features of the optical flow(using flownet2)?
Do we need to preprocess the dataset to obtain the features of vehicle's ego-motion(using orb-slam)?
Hi @sbjshxbxijs , yes please follow the instruction here to extract those features. We used the default setting of those tools when generated the features used in the paper.
Does this DoTA dataset contain optical flow features? Does it contain the features of the vehicle's ego-motion? Do we need to preprocess the data set to obtain the features of the optical flow(using flownet2)? Do we need to preprocess the dataset to obtain the features of vehicle's ego-motion(using orb-slam)?