search

tilemmpon / Singapore-Maritime-Dataset-Frames-Ground-Truth-Generation-and-Statistics

Repository for generating frames from the Singapore Maritime Dataset videos and converting the corresponding ground truth files. FInally, some basic statistics are generated.
38 stars 16 forks source link

how to use load_mat_into_csv_xml.py? #14

Open TahereSaidi opened 9 months ago

TahereSaidi commented 9 months ago

Hi, I have used this dataset and also used the repository provided by you to generate frames. But I divided the videos differently. That is, I first divided the videos into three training, evaluation and test files. As a result, there is no video with some frames in the training data and some frames in the test data.. I also put the labels of each of these three categories in separate files called train labels, evaluation labels and test labels. Now I want to convert the labels from .mat format to another format using load_mat_into_csv_xml.py file. I downloaded the load_mat_into_csv_xml.py file and used it like this:!python -i "the address of training labels" -o"the address of output", -f"the address of folder that has train and test frames" but it gives an errore: File "/Users/goly/miniconda3/lib/python3.11/site-packages/scipy/io/matlab/_mio4.py", line 179, in read_sub_array raise ValueError("Not enough bytes to read matrix '%s'; is this " ValueError: Not enough bytes to read matrix 'b'\x10\x0b\x what should i do?

tilemmpon commented 4 months ago

I apologize for the late response. Could you please let me know if you still need assistance?

TahereSaidi commented 3 months ago

no thank you. solved.

lmaple24327 commented 3 months ago

Hello, I would like to inquire about the file objects_onshore.txt, which is the tag file of the target in the video dataset On shore. I have some problems when using this tag file, saying that it is not normalized. The problem description looks like this: train WARNING ⚠️ C:\Users\admin\Desktop\ultralytics-main\datasets\MVI_1622_VIS-dataset\images\train\MVI_1622_VIS240.jpg: ignoring corrupt image/label: non-normalized or out of bounds coordinates [249.81 98.713 495.62 195.43 249.81 33.699 495.62 65.398 235.28 23.324 466.57 44.648. Do you know how to solve it? Looking forward to your reply, thank you very much.