image missing in train set jpg data dgta-seadronesee

[fabiopoiesi]

Hi and thanks for sharing this useful project. Great job!

I've downloaded the dgta-seadronesee data (jpg folder + labels + meta data) and unzipped the files. Please check the size in the respective directories.

Train

Val

The images directory of the train set is one item short.

Can you please check which file is missing?

I am going to create a script to convert the annotations in coco format, I guess I have to resize the original bboxes if I want to use them with the jpg images, right?

[C-der-Baum]

I'll have a look. I'll get back to you in a couple of hours.

[Ben93kie]

I'm currently uploading an updated version of the train part. Please let me know if there are still any issues, such as a missing image. You are right, you have to resize the original bounding boxes to 1/2 to use them with the jpg images.

[fabiopoiesi]

Thank you.

On Mon, 23 May 2022 at 22:16, Ben93kie @.***> wrote:

I'm currently uploading an updated version of the train part. Please let me know if there are still any issues, such as a missing image. You are right, you have to resize the original bounding boxes to 1/2 to use them with the jpg images.

— Reply to this email directly, view it on GitHub https://github.com/Eisbaer8/DeepGTAV/issues/4#issuecomment-1135098290, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEBXWKD672ZWCIWHIQ5BU73VLPRTPANCNFSM5WU42U6A . You are receiving this because you authored the thread.Message ID: @.***>

[fabiopoiesi]

Have you had chance to upload the dataset?

[Ben93kie]

Unfortunately, there was a connection issue and it got canceled. Uploading it again, it is at 10% and should be finished in 3-4 hours.

[Ben93kie]

I just rechecked and I have 90k images in that folder actually. I'm still uploading a new version which may resolve the problem of a potentially corrupt file. Can you send me the name of the missing or corrupt file (by comparing to the label files) if the new version does not resolve problems?

I also uploaded label-jsons that should be in coco format fyi.

[fabiopoiesi]

By running this code I obtained the following id:

fns_labels = glob.glob(os.path.join(source_dir, 'train', 'labels', '*txt'))
fns_images = glob.glob(os.path.join(source_dir, 'train', 'images', '*jpg'))
image_id_images = [os.path.basename(fn).split('.')[0] for fn in fns_images]
for fn in fns_labels:
    image_id = os.path.basename(fn).split('.')[0]
    if not image_id in image_id_images:
        print(image_id)

id -> 0020_0000168908

[fabiopoiesi]

if you share that image I can just add it to the dataset, so I don't have to re-download it

[C-der-Baum]

[fabiopoiesi]

Thanks you!