search

Duchen521 / HED-document-detection

A Pytorch Implementation of HED (Holistically-Nested Edge Detection algorithm) for Document Detection
MIT License
48 stars 6 forks source link

About the training method of the training model you provided and the training dataset #3

Open Danee-wawawa opened 4 years ago

Danee-wawawa commented 4 years ago

Hi, thank you for your job. I test the trained model you provided but the result did not look very good. In which way was the model you provided to train? （1）Train without annotated data （2）Train with annotated data And how large is the training dataset used by your model? Looking forward for your answer. Thank you.

RaoufiTech commented 2 months ago

4 years later and still no answer