Comments (7)
@Altimis looking forward get your reply as soon as possible as i have my submission to submit in coming week.
i will be grateful to you.
from confusion-matrix-for-mask-r-cnn.
Hi @nandita96, I'm sorry for the late response.
Actually, class 1 where you have 0.00% in both axes is referred to the background. The horizontal axis represents the ground-truth classes and the vertical axis represents the predicted classes. In your case, your model detected 94 objects of class 2 among the 98 objects existing in the data, the other 4 objects were identified as 3 objects of class 3 and one object of class 1 (as I said, class 1 is the background, which means that the model could not detect this object at all). The class 3 objects that exist in your images were all detected correctly (95/95 = 100%) but on the other hand, the model detected 6 other objects from the background that they are considered to be class 3 (101 detections = 95 + 6), 3 objects of class 2 and 3 objects of class 1 (the model detected 3 objects in the background as being of class 3 which are not). In conclusion, the horizontal axis represents the precisions of each class, and the vertical axis represents the recall of each class (the only reason I added the class background in the matrix is to define false positives and false negatives).
However, I still have some uncertainty about the function that computes the mAP. I will provide some time to review it.
I hope this will help.
from confusion-matrix-for-mask-r-cnn.
@Altimis
i am working on medical data, as you said vertical axis represents the recall (FN) of each class and horizontal axis represents the precisions (FP) of each class. on that basis i concluded below things, M i concluding correct ?????
will it work like that? but recall is higher than precision, i am doubt my model performance showing correct or wrong? please given your suggestion on this:
According to above confusion matrix :
class 2 has:
precision = 94.95% and recall = 95.92%
class 3 has:
precision = 94.06% and recall = 100%
from confusion-matrix-for-mask-r-cnn.
@Altimis
i am working on medical data, as you said vertical axis represents the recall (FN) of each class and horizontal axis represents the precisions (FP) of each class. on that basis i concluded below things, M i concluding correct ?????will it work like that? but recall is higher than precision, i am doubt my model performance showing correct or wrong? please given your suggestion on this:
According to above confusion matrix :
class 2 has:
precision = 94.95% and recall = 95.92%class 3 has:
precision = 94.06% and recall = 100%
In fact, what you mentioned in your first sentence is correct (vertical axis represents the recall (FN) of each class and horizontal axis represents the precisions (FP) of each class), but i think that when u came to apply this on your matrix, you confused the recall with precision (correct me if I'm wrong please).
class 2 has:
precision = 95.92% (horizontal axis) and recall = 94.95% (vertical axis)
class 3 has:
precision = 100% (horizontal axis) and recall = 94.06% (vertical axis)
from confusion-matrix-for-mask-r-cnn.
@nandita96 Did my comment help you ? So that i can close this issue.
from confusion-matrix-for-mask-r-cnn.
@Altimis , yes i able to calculate F1 score from it. Thank you so much Altimis for helping me 😊🙌... yes you can close this issue.
one thing i wanted to F1 score is Disc score , is both are same?
from confusion-matrix-for-mask-r-cnn.
In fact, I've never worked with this DISC score, but F1 is easy to calculate : the F1 score of each class i is 2 * (precision(i) * recall(i)) / (precision(i) + recall(i)).
from confusion-matrix-for-mask-r-cnn.
Related Issues (14)
- Couldnt generate confusion matrix with 2 different classes (not included background)
- Confusion Matrix Illustration HOT 1
- The confusion matrix for yolact
- Explanation of Confusion Matrix Illustration
- Adding True Negatives
- Precision-Recall curve HOT 2
- Confusion matrix HOT 6
- Confusion Matrix HOT 11
- Getting only TPs HOT 12
- computation for entire dataset HOT 1
- false positives and false negatives seem mixed up HOT 3
- AttributeError: module 'mrcnn.utils' has no attribute 'gt_pred_lists' HOT 4
- Change class number HOT 1
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 confusion-matrix-for-mask-r-cnn.