Name
Abstract | ||
|---|---|---|
•Deep metric learning can be used to classify otitis media in otoscopy images.•Manual classification is observer dependent and suffers from variable accuracy.•Triplet loss captures the variation of an unbalanced data very well.•Using triplet loss, a high accuracy of 85% was achieved.•Presence of ear wax and blurry images pose classification challenges. |
Year | DOI | Venue |
|---|---|---|
2021 | 10.1016/j.media.2021.102034 | Medical Image Analysis |
Keywords | DocType | Volume |
Otitis media,Deep metric learning,Convolutional neural network,Image classification | Journal | 71 |
ISSN | Citations | PageRank |
1361-8415 | 1 | 0.40 |
References | Authors | |
1 | 8 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Josefine Vilsbøll Sundgaard | 1 | 1 | 0.40 |
| James M. Harte | 2 | 5 | 2.23 |
| Peter Bray | 3 | 1 | 0.40 |
| Søren Laugesen | 4 | 1 | 0.40 |
| Yosuke Kamide | 5 | 1 | 0.73 |
| Chiemi Tanaka | 6 | 1 | 0.40 |
| Rasmus R. Paulsen | 7 | 88 | 20.01 |
| Anders Christensen | 8 | 1 | 1.07 |