classes colors
WeichengSong opened this issue · comments
Where do you modify the code of display colors for each category when forecasting?
Hi @WeichengSong, the colorization is done by dataset.decode_targets.
thank you，I find it.
But I dont know how to change the colors, i just want to predict 3 classes, not use the original colors types for the voc's 19 classes.
And whats meaning of the No.3 Visualize segmentation outputs codes and where to use it in your project on your github.
please help me. Thank you very much.
…---Original---
From: "Gongfan ***@***.***>
Date: Thu, Sep 1, 2022 01:04 AM
To: ***@***.***>;
Cc: ***@***.******@***.***>;
Subject: Re: [VainF/DeepLabV3Plus-Pytorch] classes colors (Issue #89)
Hi @WeichengSong, the colorization is done by dataset.decode_targets.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you were mentioned.Message ID: ***@***.***>
If you want to colorize your own datasets, you can use the VOC colormap. Here is an example of colorizing NYUv2 dataset. And a similar example can be found here in this repo.