Article Quality Measures for Image Segmentation Using Generated Images
Article Image segmentation evaluation: a survey of methods
If you have access to ground truth of masked objects (some can be found online), then one way to quantify the performance is to retrieve objects after doing superpixel segmentation and apply the Intersection over Union (IoU) method.
Intersection over Union (IoU) used for enhancing segmentation object detection you can use the morphological operation for enhancement
hope this research is useful
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.641.440&rep=rep1&type=pdf
Article Quality Measures for Image Segmentation Using Generated Images
Article Image segmentation evaluation: a survey of methods
Dear;
You can read it in the following link:
Article Image segmentation evaluation: a survey of methods
Regards
Dear;
You can read it in the following link:
Article Quality Measures for Image Segmentation Using Generated Images
Regards
You can refer to this article: "Comparison and evaluation of methods for liver segmentation from CT datasets"
Five measures are described: Volumetric Overlap Error, Relative Volume Difference, Average Symmetric Surface Distance, Root Mean Square Symmetric Surface Distance and Maximum Symmetric Surface Distance. Other popular measures include DICE and Hausdorff Distance .
Accuracy = (TP+TN)/(FN+FP+TP+TN);
Recall = TP/(TP+FN);
Precision = TP/(TP+FP);
Fmeasure = 2*TP/(2*TP+FP+FN);
MCC = (TP*TN-FP*FN)/sqrt((TP+FP)*(TP+FN)*(TN+FP)*(TN+FN));
Dice = 2*TP/(2*TP+FP+FN);
Jaccard = Dice/(2-Dice);
Specitivity = TN/(TN+FP);
Sensitivity = TP/(TP+FN);
- Badges
- Science topic
- Similar topics
- Computer Science
- Images