Saved in:
Bibliographic Details
Main Authors: Morelle, Olivier, Bisten, Justus, Wintergerst, Maximilian W. M., Finger, Robert P., Schultz, Thomas
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2501.05933
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910887881736192
author Morelle, Olivier
Bisten, Justus
Wintergerst, Maximilian W. M.
Finger, Robert P.
Schultz, Thomas
author_facet Morelle, Olivier
Bisten, Justus
Wintergerst, Maximilian W. M.
Finger, Robert P.
Schultz, Thomas
contents Weakly supervised segmentation has the potential to greatly reduce the annotation effort for training segmentation models for small structures such as hyper-reflective foci (HRF) in optical coherence tomography (OCT). However, most weakly supervised methods either involve a strong downsampling of input images, or only achieve localization at a coarse resolution, both of which are unsatisfactory for small structures. We propose a novel framework that increases the spatial resolution of a traditional attention-based Multiple Instance Learning (MIL) approach by using Layer-wise Relevance Propagation (LRP) to prompt the Segment Anything Model (SAM~2), and increases recall with iterative inference. Moreover, we demonstrate that replacing MIL with a Compact Convolutional Transformer (CCT), which adds a positional encoding, and permits an exchange of information between different regions of the OCT image, leads to a further and substantial increase in segmentation accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2501_05933
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Weakly Supervised Segmentation of Hyper-Reflective Foci with Compact Convolutional Transformers and SAM2
Morelle, Olivier
Bisten, Justus
Wintergerst, Maximilian W. M.
Finger, Robert P.
Schultz, Thomas
Computer Vision and Pattern Recognition
Weakly supervised segmentation has the potential to greatly reduce the annotation effort for training segmentation models for small structures such as hyper-reflective foci (HRF) in optical coherence tomography (OCT). However, most weakly supervised methods either involve a strong downsampling of input images, or only achieve localization at a coarse resolution, both of which are unsatisfactory for small structures. We propose a novel framework that increases the spatial resolution of a traditional attention-based Multiple Instance Learning (MIL) approach by using Layer-wise Relevance Propagation (LRP) to prompt the Segment Anything Model (SAM~2), and increases recall with iterative inference. Moreover, we demonstrate that replacing MIL with a Compact Convolutional Transformer (CCT), which adds a positional encoding, and permits an exchange of information between different regions of the OCT image, leads to a further and substantial increase in segmentation accuracy.
title Weakly Supervised Segmentation of Hyper-Reflective Foci with Compact Convolutional Transformers and SAM2
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2501.05933