search

ultralytics / xview-yolov3

xView 2018 Object Detection Challenge: YOLOv3 Training and Inference.
https://docs.ultralytics.com
GNU Affero General Public License v3.0
252 stars 57 forks source link

analysis.m does not call make_small_chips() #15

Closed sawhney-medha closed 5 years ago

sawhney-medha commented 5 years ago

The matlab code analysis.m does not call make_small_chips() function anywhere. So to run the code, is the function call supposed to be uncommented or run without it?

glenn-jocher commented 5 years ago

make_small_chips() creates smaller images from each bounding box to pass to a second stage classifier.

Be advised that the https://github.com/ultralytics/xview-yolov3 repository is not under active development anymore. We recommend you use https://github.com/ultralytics/yolov3 instead, our main YOLOv3 repository.

sawhney-medha commented 5 years ago

@glenn-jocher Thank you for the prompt reply. I understand that it is not being updated but my work is specific to xview data hence I want to understand the working of this. So if I want to train on my custom data or train on lesser classes do i need to call make_small_chips() while running analysis.m script or would the chips used for training will be taken care by datasets.py? Do i need to use make_small_chips() for creating a targets_c12.mat for my subset of xview classes?

glenn-jocher commented 5 years ago

@sawhney-medha datasets.py selects chips from the available training data, you don't need make_small_chips(). make_small_chips() makes small chips from individual bounding boxes, some as small as 5x5 pixels for example, for a second stage classifier.

sawhney-medha commented 5 years ago

Okay. Thanks a lot for the help!

glenn-jocher commented 10 months ago

@sawhney-medha you're welcome! If you have any other questions, feel free to ask. Good luck with your project!