Learning with partial labels: an exploratory study of data scarcity for medical images
<p>The data-driven nature of deep learning models is often associated with the use of large-scale fully labeled datasets. However, the high cost of annotation often makes these datasets unavailable for standard supervised learning. This thesis explores an emerging learning paradigm called part...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2023
|
| Subjects: |