Topology-Preserving Deep Image Segmentation
Problem
Segmentation algorithms make topological errors, such as broken connections, on fine-scale structures in images.
Why it matters
Topological errors undermine reliability of automated image analysis in biomedical and natural imaging.
What the AI did
Designed a differentiable topology-preserving loss function enforcing the same Betti number as ground truth, incorporated into end-to-end training of a deep neural network.
Results
Improved Betti number error, Adjusted Rand Index and Variation of Information; no numeric values stated.
Impact
Improved topological correctness of segmentation across natural and biomedical datasets per authors.
Record
| Venue | arXiv (Cornell University) |
|---|---|
| Published | 2019 |
| Work type | preprint |
| Authors | Hu, Xiaoling, Fuxin Li, Dimitris Samaras, Chao Chen |
| Citations | 114 (OpenAlex, at import time) |
| References | 38 |
| Model or method | Deep neural network with topology-preserving (Betti number) loss |
| Organization | not stated in source |
| Primary source | https://doi.org/10.48550/arxiv.1906.05404 |
| Open access | green |
| Keywords | image segmentation, topology, Betti number, deep learning, loss function |
Importance rating
real but narrower advance
Topological errors in segmentation affect reliability of automated image analysis across biomedical and natural imaging; a useful methodological advance but no stated application scale.
More in Software & Coding
| # | Paper | Year | Citations | Importance |
|---|---|---|---|---|
| 1 | ImageNet classification with deep convolutional neural networks | 2017 | 106,345 |
8
|
| 2 | Deep Residual Learning for Image Recognition | 2016 | 221,134 |
8
|
| 3 | Faster sorting algorithms discovered using deep reinforcement learning | 2023 | 175 |
8
|
| 4 | A deep learning‐based approach for software vulnerability detection using code metrics | 2022 | 42 |
7
|
| 5 | Deep Learning-Based Attack Detection and Classification in Android Devices | 2023 | 41 |
7
|