ZJULearning / pixel_link

Implementation of our paper 'PixelLink: Detecting Scene Text via Instance Segmentation' in AAAI2018
MIT License
768 stars 255 forks source link

Where is base training model? #160

Closed famunir closed 4 years ago

famunir commented 4 years ago

Hello, Thank you for your efforts. Can you please tell me which base model is used for training? Do I have to download vgg16 or is your pretrained model (which is provided in README.md) used as a base model for further training? I have some issue with the result and want to train more for by images. Any help from you guys will be appreciated. Thank again! @dengdan @comzyh @bengxy @fc731097343 @ewanlee

famunir commented 4 years ago

Solved! I am using this base model for further training. PixelLink + VGG16 2s provided here