Comments (4)
SorourMo
commented on July 28, 2024
Hi.
The evaluation metrics described in Cloud-net are calculated first over each test image. Then they are averaged over the total number of test images. Like this:
Jaccard of the first test image: Jaccard1 = tp1/(tp1+fp1+fn1)
Jaccard of the second test image: Jaccard2 = tp2/(tp2+fp2+fn2)
.
.
Total Jaccard = (Jaccard1+Jaccard2+...+JaccardM)/M
where M is the total num of images.
In Cloud-net+, tp, fp, fn, and tn are measured for each target class and then metrics are calculated as mentioned in the "evaluation metrics" section of the paper. Like this:
Total Jaccard = (tp1 + tp2+ ... + tpM)/((tp1 + tp2+ ... + tpM)+(fp1 + fp2+ ... + fpM)+(fn1 + fn2+ ... + fnM))
That is the same for other metrics too.
Cloud-Net+ code will be available soon.
from 38-cloud-a-cloud-segmentation-dataset.
Silengcewang
commented on July 28, 2024
Thank you for your reply.
Because that we usually use the first evaluatuion metrics and the second one seems have a better performance.So i'm very confused.Whatever thank you for your reply again and i will try the to metrics to find the difference.
…------------------ 原始邮件 ------------------
发件人: "SorourMo/38-Cloud-A-Cloud-Segmentation-Dataset" <[email protected]>;
发送时间: 2020年11月30日(星期一) 凌晨0:02
收件人: "SorourMo/38-Cloud-A-Cloud-Segmentation-Dataset"<[email protected]>;
抄送: "鼬手边是佐手"<[email protected]>;"Author"<[email protected]>;
主题: Re: [SorourMo/38-Cloud-A-Cloud-Segmentation-Dataset] Why are the two evaluation indicators very different (#7)
Hi.
The evaluation metrics described in Cloud-net are calculated first over each test image. Then they are averaged over the total number of test images. Like this:
Jaccard of the first test image: Jaccard1 = tp1/(tp1+fp1+fn1)
Jaccard of the second test image: Jaccard2 = tp2/(tp2+fp2+fn2)
.
.
Total Jaccard = (Jaccard1+Jaccard2+...)/M
where M is the total num of images.
In Cloud-net+, tp, fp, fn, and tn are measured for each target class and then metrics are calculated as mentioned in the "evaluation metrics" section of the paper. Like this:
Total Jaccard = (tp1 + tp2+ ... + tpM)/((tp1 + tp2+ ... + tpM)+(fp1 + fp2+ ... + fpM)+(fn1 + fn2+ ... + fnM))
That is the same for other metrics too.
Cloud-Net+ code will be available soon.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub, or unsubscribe.
from 38-cloud-a-cloud-segmentation-dataset.
Silengcewang
commented on July 28, 2024
Hello.
Sorry to disturb you again.But I really can't understand why the evaluation metrics of the same network can improve so much in your three papers:FCN、Cloud-net and Cloude-net++.If you use the same code wirted by matlab or use some special tricks.
Really,thank you for your reply!
…------------------ 原始邮件 ------------------
发件人: "SorourMo/38-Cloud-A-Cloud-Segmentation-Dataset" <[email protected]>;
发送时间: 2020年11月30日(星期一) 凌晨0:02
收件人: "SorourMo/38-Cloud-A-Cloud-Segmentation-Dataset"<[email protected]>;
抄送: "鼬手边是佐手"<[email protected]>;"Author"<[email protected]>;
主题: Re: [SorourMo/38-Cloud-A-Cloud-Segmentation-Dataset] Why are the two evaluation indicators very different (#7)
Hi.
The evaluation metrics described in Cloud-net are calculated first over each test image. Then they are averaged over the total number of test images. Like this:
Jaccard of the first test image: Jaccard1 = tp1/(tp1+fp1+fn1)
Jaccard of the second test image: Jaccard2 = tp2/(tp2+fp2+fn2)
.
.
Total Jaccard = (Jaccard1+Jaccard2+...)/M
where M is the total num of images.
In Cloud-net+, tp, fp, fn, and tn are measured for each target class and then metrics are calculated as mentioned in the "evaluation metrics" section of the paper. Like this:
Total Jaccard = (tp1 + tp2+ ... + tpM)/((tp1 + tp2+ ... + tpM)+(fp1 + fp2+ ... + fpM)+(fn1 + fn2+ ... + fnM))
That is the same for other metrics too.
Cloud-Net+ code will be available soon.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub, or unsubscribe.
from 38-cloud-a-cloud-segmentation-dataset.
SorourMo
commented on July 28, 2024
There is no trick. The Matlab code provided here are only for the FCN and Cloud-Net papers. As I said before (here and in the papers), this code calculates metrics for each image and then averages those metrics over 20 scenes of 38-Cloud test set.
However, if you want to calculate metrics over each class, and not each image, (same as what've done in Cloud-Net+ paper), you need to write your own code to do so.
from 38-cloud-a-cloud-segmentation-dataset.
Related Issues (6)
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 38-cloud-a-cloud-segmentation-dataset.