lcy0604 / CTRNet

This repository is the implementation of "Don't Forget Me: Accurate Background Recovery for Text Removal via Modeling Local-Global Context".

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

RTV方法生成structure-im时间过长问题

hijupiter opened this issue · comments

请问老师,训练的时候LCG的模块需要输入structure-im图像。
structure-im图像由RTV方法生成,非常耗时和耗内存。
模型推理的时候也需要structure-im,会造成推理时间过长,请问老师有没有好的解决方式?

I can give you some advice.
Actually you can just don't use structure-im image. It won't make much difference to your results.

I can give you some advice. Actually you can just don't use structure-im image. It won't make much difference to your results.

hi,what is your advice? can you explain in detail

Structure-im has little effect on the result, so you don't have to use structure-im. Just input the mask and the origin image.