RankSEG: A Consistent Ranking-based Framework for Segmentation

by   Ben Dai, et al.

Segmentation has emerged as a fundamental field of computer vision and natural language processing, which assigns a label to every pixel/feature to extract regions of interest from an image/text. To evaluate the performance of segmentation, the Dice and IoU metrics are used to measure the degree of overlap between the ground truth and the predicted segmentation. In this paper, we establish a theoretical foundation of segmentation with respect to the Dice/IoU metrics, including the Bayes rule and Dice/IoU-calibration, analogous to classification-calibration or Fisher consistency in classification. We prove that the existing thresholding-based framework with most operating losses are not consistent with respect to the Dice/IoU metrics, and thus may lead to a suboptimal solution. To address this pitfall, we propose a novel consistent ranking-based framework, namely RankDice/RankIoU, inspired by plug-in rules of the Bayes segmentation rule. Three numerical algorithms with GPU parallel execution are developed to implement the proposed framework in large-scale and high-dimensional segmentation. We study statistical properties of the proposed framework. We show it is Dice-/IoU-calibrated, and its excess risk bounds and the rate of convergence are also provided. The numerical effectiveness of RankDice/mRankDice is demonstrated in various simulated examples and Fine-annotated CityScapes and Pascal VOC datasets with state-of-the-art deep learning architectures.


page 18

page 22


Optimization with soft Dice can lead to a volumetric bias

Segmentation is a fundamental task in medical image analysis. The clinic...

On the Consistency of Top-k Surrogate Losses

The top-k error is often employed to evaluate performance for challengin...

A 3D Coarse-to-Fine Framework for Automatic Pancreas Segmentation

In this paper, we adopt 3D CNNs to segment the pancreas in CT images. Al...

A Unifying Theory of Distance from Calibration

We study the fundamental question of how to define and measure the dista...

The hidden label-marginal biases of segmentation losses

Most segmentation losses are arguably variants of the Cross-Entropy (CE)...

Automated segmentation of intracranial hemorrhages from 3D CT

Intracranial hemorrhage segmentation challenge (INSTANCE 2022) offers a ...

Please sign up or login with your details

Forgot password? Click here to reset