Unsupervised Landmark detection
Title: |
Unsupervised Landmark detection |
DNr: |
Berzelius-2023-83 |
Project Type: |
LiU Berzelius |
Principal Investigator: |
Markus Ekvall <marekv@kth.se> |
Affiliation: |
Kungliga Tekniska högskolan |
Duration: |
2023-03-30 – 2023-10-01 |
Classification: |
10203 |
Keywords: |
|
Abstract
Spatial landmarks are useful for comparing histological heterogeneity between samples or sites, tracking regions of interest in microscopy, and registering tissue samples in a common coordinate framework. While unsupervised landmark detection has been studied previously, current methods are unsuitable for histological image data as they typically require a large number of images and do not account for non-linear deformations between tissue sections. We address these limitations by proposing a novel method for landmark detection and registration using neural-network-guided thin plate splines. We evaluate the proposed method on a wide range of datasets and show that it outperforms existing approaches in terms of both accuracy and stability.