tensorflow-object-contour-detection
tensorflow-object-contour-detection copied to clipboard
Dataset to be used for training
Hi, Thanks for sharing the code. I am little unsure about the dataset that shoud be used for training. Also, how it should be pre-processed. The paper says that PASCAL VOC 2012 has been used with some CRF pre-processing on the polygon annotations provided for the segmentation.
I have downloaded PASCAL VOC 2012 from http://host.robots.ox.ac.uk/pascal/VOC/voc2012/#data . The images folder contains 17000 images. How should I generate the ground truth from this?
Hoping to get some response.
Thanks
The ground-truth is already within the PASCAL VOC dataset. Go to VOC2012> Segmentation Class folder. You can see all the ground-truth masks over there.
The ground-truth is already within the PASCAL VOC dataset. Go to VOC2012> Segmentation Class folder. You can see all the ground-truth masks over there.
Can you please tell me how did you exactly train the model? Even I tried using the Segmentation Class folder as the ground truth values.