# Reverse Classification Accuracy: Predicting Segmentation Performance in the Absence of Ground Truth

arXiv: https://arxiv.org/abs/1702.03407 · HTML: https://mlsanity.com/paper/1702.03407/

## Key facts

- Cited by 35 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 Feb 2017.

## Built on by (most cited first, 30 of 35)

- [Robust machine learning segmentation for large-scale analysis of heterogeneous clinical brain MRI datasets](https://mlsanity.com/paper/2209.02032/) (2022)
- [CheXmask: a large-scale dataset of anatomical segmentation masks for multi-center chest x-ray images](https://mlsanity.com/paper/2307.03293/) (2023)
- [Comparative Benchmarking of Failure Detection Methods in Medical Image Segmentation: Unveiling the Role of Confidence Aggregation](https://mlsanity.com/paper/2406.03323/) (2024)
- [Estimating Model Performance under Domain Shifts with Class-Specific Confidence Scores](https://mlsanity.com/paper/2207.09957/) (2022)
- [FSNet: A Failure Detection Framework for Semantic Segmentation](https://mlsanity.com/paper/2108.08748/) (2021)
- [Conformal Performance Range Prediction for Segmentation Output Quality Control](https://mlsanity.com/paper/2407.13307/) (2024)
- [Efficient Model Monitoring for Quality Control in Cardiac Image Segmentation](https://mlsanity.com/paper/2104.05533/) (2021)
- [Deep Quality Estimation: Creating Surrogate Models for Human Quality Ratings](https://mlsanity.com/paper/2205.10355/) (2022)
- [U-Netmer: U-Net meets Transformer for medical image segmentation](https://mlsanity.com/paper/2304.01401/) (2023)
- [Mask of truth: model sensitivity to unexpected regions of medical images](https://mlsanity.com/paper/2412.04030/) (2024)
- [Expectation-Maximization as the Engine of Scalable Medical Intelligence](https://mlsanity.com/paper/2501.03410/) (2025)
- [Promises and pitfalls of deep neural networks in neuroimaging-based psychiatric research](https://mlsanity.com/paper/2301.08525/) (2023)
- [SUDO: a framework for evaluating clinical artificial intelligence systems without ground-truth annotations](https://mlsanity.com/paper/2403.17011/) (2024)
- [ConfIC-RCA: Statistically Grounded Efficient Estimation of Segmentation Quality](https://mlsanity.com/paper/2503.04522/) (2025)
- [Automatic identification of segmentation errors for radiotherapy using geometric learning](https://mlsanity.com/paper/2206.13317/) (2022)
- [Towards Ground-truth-free Evaluation of Any Segmentation in Medical Images](https://mlsanity.com/paper/2409.14874/) (2024)
- [Deep and Statistical Learning in Biomedical Imaging: State of the Art in 3D MRI Brain Tumor Segmentation](https://mlsanity.com/paper/2103.05529/) (2021)
- [Open Challenges on Fairness of Artificial Intelligence in Medical Imaging Applications](https://mlsanity.com/paper/2407.16953/) (2024)
- [Uncertainty-Aware Segmentation Quality Prediction via Deep Learning Bayesian Modeling: Comprehensive Evaluation and Interpretation on Skin Cancer and Liver Segmentation](https://mlsanity.com/paper/2508.01460/) (2025)
- [Hardware-aware Real-time Myocardial Segmentation Quality Control in Contrast Echocardiography](https://mlsanity.com/paper/2109.06909/) (2021)
- [Automatic quality control framework for more reliable integration of machine learning-based image segmentation into medical workflows](https://mlsanity.com/paper/2112.03277/) (2021)
- [Unsupervised bias discovery in medical image segmentation](https://mlsanity.com/paper/2309.00451/) (2023)
- [Cross-Model Agreement as a Deployment-Time Reliability Signal for Automatic Polyp Segmentation](https://mlsanity.com/paper/2609.10495/) (2026)
- [Quality monitoring of federated Covid-19 lesion segmentation](https://mlsanity.com/paper/2112.08974/) (2021)
- [TMS-Net: A Segmentation Network Coupled With A Run-time Quality Control Method For Robust Cardiac Image Segmentation](https://mlsanity.com/paper/2212.10877/) (2022)
- [RAISE -- Radiology AI Safety, an End-to-end lifecycle approach](https://mlsanity.com/paper/2311.14570/) (2023)
- [Machine Learning Techniques for MRI Data Processing at Expanding Scale](https://mlsanity.com/paper/2404.14326/) (2024)
- [QCResUNet: Joint Subject-level and Voxel-level Segmentation Quality Prediction](https://mlsanity.com/paper/2412.07156/) (2024)
- [Unsupervised Source-Free Ranking of Biomedical Segmentation Models Under Distribution Shift](https://mlsanity.com/paper/2503.00450/) (2025)
- [Performance Estimation for Supervised Medical Image Segmentation Models on Unlabeled Data Using UniverSeg](https://mlsanity.com/paper/2504.15667/) (2025)

## Abstract (arXiv)

When integrating computational tools such as automatic segmentation into clinical practice, it is of utmost importance to be able to assess the level of accuracy on new data, and in particular, to detect when an automatic method fails. However, this is difficult to achieve due to absence of ground truth. Segmentation accuracy on clinical data might be different from what is found through cross-validation because validation data is often used during incremental method development, which can lead to overfitting and unrealistic performance expectations. Before deployment, performance is quantified using different metrics, for which the predicted segmentation is compared to a reference segmentation, often obtained manually by an expert. But little is known about the real performance after deployment when a reference is unavailable. In this paper, we introduce the concept of reverse classification accuracy (RCA) as a framework for predicting the performance of a segmentation method on new data. In RCA we take the predicted segmentation from a new image to train a reverse classifier which is evaluated on a set of reference images with available ground truth. The hypothesis is that if the predicted segmentation is of good quality, then the reverse classifier will perform well on at least some of the reference images. We validate our approach on multi-organ segmentation with different classifiers and segmentation methods. Our results indicate that it is indeed possible to predict the quality of individual segmentations, in the absence of ground truth. Thus, RCA is ideal for integration into automatic processing pipelines in clinical routine and as part of large-scale image analysis studies.