Surgical phase and instrument recognition: how to identify appropriate dataset splits
Machine learning approaches can only be reliably evaluated if training, validation, and test data splits are representative and not affected by the absence of classes. Surgical workflow and instrument recognition are two tasks that are complicated in this manner, because of heavy data imbalances res...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
29 January 2024
|
| In: |
International journal of computer assisted radiology and surgery
Year: 2024, Volume: 19, Issue: 4, Pages: 699-711 |
| ISSN: | 1861-6429 |
| DOI: | 10.1007/s11548-024-03063-9 |
| Online Access: | Verlag, kostenfrei, Volltext: https://doi.org/10.1007/s11548-024-03063-9 Verlag, kostenfrei, Volltext: https://link.springer.com/article/10.1007/s11548-024-03063-9 
|
| Author Notes: | Georgii Kostiuchik, Lalith Sharan, Benedikt Mayer, Ivo Wolf, Bernhard Preim, Sandy Engelhardt |
| Summary: | Machine learning approaches can only be reliably evaluated if training, validation, and test data splits are representative and not affected by the absence of classes. Surgical workflow and instrument recognition are two tasks that are complicated in this manner, because of heavy data imbalances resulting from different length of phases and their potential erratic occurrences. Furthermore, sub-properties like instrument (co-)occurrence are usually not particularly considered when defining the split. |
|---|---|
| Item Description: | Gesehen am 05.11.2024 |
| Physical Description: | Online Resource |
| ISSN: | 1861-6429 |
| DOI: | 10.1007/s11548-024-03063-9 |