How to calculate The classification error rate (E) of iris recognition
1 回表示 (過去 30 日間)
古いコメントを表示
Hello, How to calculate The classification error rate (E) of iris recognition in this article by using MATLAB? http://www.di.ubi.pt/~hugomcp/doc/ProencaAlexandreNICE1BTAS2007.pdf A lot of articles used this to calculate the segmentation accuracy. I want to compare the segmentation accuracy of two iris images (ground truth & enhanced image) TQ
0 件のコメント
回答 (1 件)
Image Analyst
2015 年 8 月 29 日
What's the question? You should know if it identified the correct individual or not. Whatever rates you want to compute can be determined by the true positive, true negative, false positive, and false negative (TP, TN, FP, FN) numbers. From that you can construct the ROC curve.
3 件のコメント
Image Analyst
2015 年 8 月 29 日
OK, they're not defining ground truth by whether it identified the correct individual based on the iris, but by whether their algorithm matches up pixel for pixel with some other binary image that they call "ground truth" which is possibly some binary image created by someone hand-drawing the boundaries.
So, is there something about the xor() function that you don't understand? Just use it like this:
differentPixels = xor(thisImage, groundTruthImage);
fractionDiffPixels = sum(differentPixels(:)) / numel(differentPixels);
参考
カテゴリ
Help Center および File Exchange で Recognition, Object Detection, and Semantic Segmentation についてさらに検索
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!