Comments (2)
It’s very likely that the processing script is not suitable for your dataset. You need to figure out the logics for processing RUGD and apply to your dataset. I could not figure out the issue based on the information you provided.
If the first img is the ID img, it should not have white color as the value is vary large. So it must be some bugs directly using rugd script for your dataset. You might have to go through your code and figure out the specific bugs.
I’m closing the issue since it’s not about processing of rugd or rellis dataset.
from ganav-offroad.
Thanks for your reply!
Yes, it is the ID img, it seems that all the background are painted in white in that image.
I think I found the problem, in the '( )_relabel6.py' file , at line 35:
color_id[tuple([0, 0, 0])] = 255
This line of code will change the id of the class with RGB value (0, 0, 0) to 255, I'm not sure what is this line of code used for.
When I run the original 'rugd_relabel6.py' in colab, I added two print() function to see the value of the parameter 'color_id', and it turns out after running the code color_id[tuple([0, 0, 0])] = 255
, an extra value (0, 0, 0): 255
is added to the parameter 'color_id':
PALETTE = [ [ 108, 64, 20 ], [ 255, 229, 204 ],[ 0, 102, 0 ],[ 0, 255, 0 ],
[ 0, 153, 153 ],[ 0, 128, 255 ],[ 0, 0, 255 ],[ 255, 255, 0 ],[ 255, 0, 127 ],
[ 64, 64, 64 ],[ 255, 128, 0 ],[ 255, 0, 0 ],[ 153, 76, 0 ],[ 102, 102, 0 ],
[ 102, 0, 0 ],[ 0, 255, 128 ],[ 204, 153, 255 ],[ 102, 0, 204 ],[ 255, 153, 204 ],
[ 0, 102, 102 ],[ 153, 204, 255 ],[ 102, 255, 255 ],[ 101, 101, 11 ],[ 114, 85, 47 ] ]
color_id = {tuple(c):i for i, c in enumerate(PALETTE)}
print(color_id)
color_id[tuple([0, 0, 0])] = 255
print(color_id)
result:
{(108, 64, 20): 0, (255, 229, 204): 1, (0, 102, 0): 2, (0, 255, 0): 3, (0, 153, 153): 4, (0, 128, 255): 5, (0, 0, 255): 6, (255, 255, 0): 7, (255, 0, 127): 8, (64, 64, 64): 9, (255, 128, 0): 10, (255, 0, 0): 11, (153, 76, 0): 12, (102, 102, 0): 13, (102, 0, 0): 14, (0, 255, 128): 15, (204, 153, 255): 16, (102, 0, 204): 17, (255, 153, 204): 18, (0, 102, 102): 19, (153, 204, 255): 20, (102, 255, 255): 21, (101, 101, 11): 22, (114, 85, 47): 23}
{(108, 64, 20): 0, (255, 229, 204): 1, (0, 102, 0): 2, (0, 255, 0): 3, (0, 153, 153): 4, (0, 128, 255): 5, (0, 0, 255): 6, (255, 255, 0): 7, (255, 0, 127): 8, (64, 64, 64): 9, (255, 128, 0): 10, (255, 0, 0): 11, (153, 76, 0): 12, (102, 102, 0): 13, (102, 0, 0): 14, (0, 255, 128): 15, (204, 153, 255): 16, (102, 0, 204): 17, (255, 153, 204): 18, (0, 102, 102): 19, (153, 204, 255): 20, (102, 255, 255): 21, (101, 101, 11): 22, (114, 85, 47): 23, (0, 0, 0): 255}
So I tried to comment out this line of code for my own dataset, and after pre-processing the processed image looked like this:
So there's no white color on the iD img now, since there are just 6 classes represented by RGB value from 0~5, all the pixel are in low RGB values, so theoretically it should be correct. So I think may be this line of code is not nessary: color_id[tuple([0, 0, 0])] = 255
.
from ganav-offroad.
Related Issues (20)
- Check point file HOT 10
- about the training loss become nan HOT 6
- AssertionError HOT 8
- Enquiry regarding Performance of Training HOT 14
- Error in ONNX conversion HOT 10
- how to get segmentation cost map from segmentation results? HOT 1
- Conversion of RUGD and Rellis Datasets To Rugd6 Group & Training HOT 4
- Is the onnx file generated by pytorch2onnx.py quantized? HOT 1
- Welcome update to OpenMMLab 2.0 HOT 1
- Use pretrained ResNet50 model (backbone) on ImageNet.
- L1? L2? L3? HOT 2
- checkpoint error HOT 4
- Output image of the model HOT 5
- ModuleNotFoundError: No module named 'mmcv._ext' HOT 4
- Change of backbone HOT 2
- Training on GOOSE Dataset HOT 3
- rellis folder structure
- output quality HOT 2
- User Guide / Usage HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from ganav-offroad.