YCB-V dataset

[DanieleBertagnoli]

Hi I was trying to train this model using YCB-V dataset (in order to do some tests for the https://github.com/aau-cns/poet model). Can you link me the YCB version you used for training this model? I noticed that in datasets.py you look for labels and images files that are not present in the standard YCB dataset.

Best regards, Daniele

[tgjantos]

Hi Daniele,

I am using the YCB-V dataset as provided by the BOP Challenge. However, YOLO requires a specific annotation style that is not present in the original YCB-V dataset. YOLO requires text files containing the paths to all training/validation/test images. Moreover, for each training image, e.g. image_001.jpg, you need a text file, i.e. image_001.txt, that contains the annotation information: Obj class, bounding box information (x_c, y_c, w, h normalized by the image size). I would suggest that you write a script, that uses the provided annotation information from YCB-V and traverses the whole dataset to create the required annotation files.

Best,
Thomas