Open MuhammadAsadJaved opened 3 years ago
the mat file can be loaded like
import scipy.io
mat = scipy.io.loadmat('/content/Drone_frames/img/ground_truth.mat')
but yes, seems to be some issue with the number of annotations and number of images, i am confused as well
but yes, seems to be some issue with the number of annotations and number of images, i am confused as well
hello sir, i also find the same problem after downloading the USC drone dataset, would you mind re-uploading the package of your USC drone dataset's annotations or email a separate copy to me? Thanks a lot! 819738528@qq.com
Hello sir, facing the problems with the annotations being less than the images, please can you share the correct annotations.
Did you solve the problem with the annotations??
Hi Sir, Thanks for your great work and contributions. I have a question about Annotations.
1- For usc drone set there are several subfolders for example Drone8 which contains 3200 images, however, the "groundtruth_rect.txt" file contains only 214 annotations. All other folders are the same. Can you explain the annotations more? or are you using only every 15th frame from these sets and annotations are also given for only every 15th frame? 2- For youtube the annotations are given in the .mat format, can you please provide in .txt, .csv or .xml file?
3-There are no annotations for the Thermal_data, the only detection results contain annotations. what about the other parts? Can you upload or send all annotations?
Looking forward to hearing from you. Asad Javed asadjavedgujar@gmail.com