| Title | Improving Domain Generalization in Segmentation Models with Neural Style Transfer |
| Publication Type | Conference Paper |
| Year of Publication | 2021 |
| Authors | Kline, Timothy L. |
| Conference Name | 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI) |
| Keywords | Biological system modeling, convolutional neural networks, image segmentation, Kernel, magnetic resonance imaging, Metrics, neural style transfer, polycystic kidney disease, pubcrawl, resilience, Resiliency, Scalability, semantic segmentation, Shape, Task Analysis, Training data |
| Abstract | Generalizing automated medical image segmentation methods to new image domains is inherently difficult. We have previously developed a number of automated segmentation methods that perform at the level of human readers on images acquired under similar conditions to the original training data. We are interested in exploring techniques that will improve model generalization to new imaging domains. In this study we explore a method to limit the inherent bias of these models to intensity and textural information. Using a dataset of 100 T2-weighted MR images with fat-saturation, and 100 T2-weighted MR images without fat-saturation, we explore the use of neural style transfer to induce shape preference and improve model performance on the task of segmenting the kidneys in patients affected by polycystic kidney disease. We find that using neural style transfer images improves the average dice value by 0.2. In addition, visualizing individual network kernel responses highlights a drastic difference in the optimized networks. Biasing models to invoke shape preference is a promising approach to create methods that are more closely aligned with human perception. |
| DOI | 10.1109/ISBI48211.2021.9433968 |
| Citation Key | kline_improving_2021 |
Improving Domain Generalization in Segmentation Models with Neural Style Transfer