search

GeorgeSeif / Semantic-Segmentation-Suite

Semantic Segmentation Suite in TensorFlow. Implement, train, and test new Semantic Segmentation models easily!
2.51k stars 880 forks source link

Input normalization #134

Closed zhly0 closed 5 years ago

zhly0 commented 5 years ago

Hi,GeorgeSeif: This is not a issue,but a question:

  1. In your code,images were normalised to 0-1,but if i use MobilenetV2 as frontend,is the MobilenetV2 were pretrained under the same preprocessing,which means the input image were normalised into 0-1 too?And the default input size is 512x512,is the same to the pretrained MobilenetV2's input size?
  2. And the MobilenetV2 model your provide,which dataset you pretrained on? Thanks!
GeorgeSeif commented 5 years ago

Please fill out the issue template so I can give you a more accurate answer :)