You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for this great work! I am currently studying the code related to establishing a dataset, and I have encountered some confusion regarding the formula used to calculate the crop parameters, particularly the variables cc and crop_width. These parameters seem to appear rather abruptly, and I am struggling to understand their derivation within the dataset creation process.
Given the importance of this operation to this work, I would greatly appreciate it if you could kindly share any relevant resources that might explain how these parameters are calculated and utilized. Your guidance would be invaluable in helping me move forward with this project.
Additionally, in the line of code crop_width = 2 * s * (bbox[2] - bbox[0]) / length, you are using (bbox[2] - bbox[0]), which represents the width of the original bounding box (bbx), to calculate the width of the cropped bounding box. However, since the original bounding box has been transformed into a square and subsequently jittered, would it be more appropriate to use the width of the jittered bounding box, calculated as bbox_jitter[2] - bbox_jitter[0] for this operation?
The text was updated successfully, but these errors were encountered:
Thanks for this great work! I am currently studying the code related to establishing a dataset, and I have encountered some confusion regarding the formula used to calculate the crop parameters, particularly the variables
cc
andcrop_width
. These parameters seem to appear rather abruptly, and I am struggling to understand their derivation within the dataset creation process.Given the importance of this operation to this work, I would greatly appreciate it if you could kindly share any relevant resources that might explain how these parameters are calculated and utilized. Your guidance would be invaluable in helping me move forward with this project.
Additionally, in the line of code
crop_width = 2 * s * (bbox[2] - bbox[0]) / length
, you are using(bbox[2] - bbox[0])
, which represents the width of the original bounding box (bbx), to calculate the width of the cropped bounding box. However, since the original bounding box has been transformed into a square and subsequently jittered, would it be more appropriate to use the width of the jittered bounding box, calculated asbbox_jitter[2] - bbox_jitter[0]
for this operation?The text was updated successfully, but these errors were encountered: