# Self-Learning to Detect and Segment Cysts in Lung CT Images without Manual Annotation

arXiv: https://arxiv.org/abs/1801.08486 · HTML: https://mlsanity.com/paper/1801.08486/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 Jan 2018.

## Built on by (most cited first, 6 of 6)

- [Inter Extreme Points Geodesics for End-to-End Weakly Supervised Image Segmentation](https://mlsanity.com/paper/2107.00583/) (2021)
- [Devil is in the Queries: Advancing Mask Transformers for Real-world Medical Image Segmentation and Out-of-Distribution Localization](https://mlsanity.com/paper/2304.00212/) (2023)
- [Label-free segmentation from cardiac ultrasound using self-supervised learning](https://mlsanity.com/paper/2210.04979/) (2022)
- [Cluster-Induced Mask Transformers for Effective Opportunistic Gastric Cancer Screening on Non-contrast CT Scans](https://mlsanity.com/paper/2307.04525/) (2023)
- [Boosting Medical Image-based Cancer Detection via Text-guided Supervision from Reports](https://mlsanity.com/paper/2405.14230/) (2024)
- [External Attention Assisted Multi-Phase Splenic Vascular Injury Segmentation with Limited Data](https://mlsanity.com/paper/2201.00942/) (2022)

## Abstract (arXiv)

Image segmentation is a fundamental problem in medical image analysis. In recent years, deep neural networks achieve impressive performances on many medical image segmentation tasks by supervised learning on large manually annotated data. However, expert annotations on big medical datasets are tedious, expensive or sometimes unavailable. Weakly supervised learning could reduce the effort for annotation but still required certain amounts of expertise. Recently, deep learning shows a potential to produce more accurate predictions than the original erroneous labels. Inspired by this, we introduce a very weakly supervised learning method, for cystic lesion detection and segmentation in lung CT images, without any manual annotation. Our method works in a self-learning manner, where segmentation generated in previous steps (first by unsupervised segmentation then by neural networks) is used as ground truth for the next level of network learning. Experiments on a cystic lung lesion dataset show that the deep learning could perform better than the initial unsupervised annotation, and progressively improve itself after self-learning.