mlsanityData to 7 Oct 2026

Image quality assessment by overlapping task-specific and task-agnostic measures: application to prostate multiparametric MR images for cancer segmentation

arXiv 2202.09798 ↗First posted 20 Feb 2022eess.IV cs.CVcited by 8

Not on the map. Why?

Abstract (arXiv)

Image quality assessment (IQA) in medical imaging can be used to ensure that downstream clinical tasks can be reliably performed. Quantifying the impact of an image on the specific target tasks, also named as task amenability, is needed. A task-specific IQA has recently been proposed to learn an image-amenability-predicting controller simultaneously with a target task predictor. This allows for the trained IQA controller to measure the impact an image has on the target task performance, when this task is performed using the predictor, e.g. segmentation and classification neural networks in modern clinical applications. In this work, we propose an extension to this task-specific IQA approach, by adding a task-agnostic IQA based on auto-encoding as the target task. Analysing the intersection between low-quality images, deemed by both the task-specific and task-agnostic IQA, may help to differentiate the underpinning factors that caused the poor target task performance. For example, common imaging artefacts may not adversely affect the target task, which would lead to a low task-agnostic quality and a high task-specific quality, whilst individual cases considered clinically challenging, which can not be improved by better imaging equipment or protocols, is likely to result in a high task-agnostic quality but a low task-specific quality. We first describe a flexible reward shaping strategy which allows for the adjustment of weighting between task-agnostic and task-specific quality scoring. Furthermore, we evaluate the proposed algorithm using a clinically challenging target task of prostate tumour segmentation on multiparametric magnetic resonance (mpMR) images, from 850 patients. The proposed reward shaping strategy, with appropriately weighted task-specific and task-agnostic qualities, successfully identified samples that need re-acquisition due to defected imaging process.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 17 of its 81 references that are in the map. First posted on arXiv 20 Feb 2022.

Builds on

17 of 81 references.

References · 17

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  2. 2017Proximal Policy Optimization Algorithmscited by 14,441
  3. 2015Continuous control with deep reinforcement learningcited by 3,279
  4. 2016Neural Architecture Search with Reinforcement Learningcited by 1,489
  5. 20163D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotationcited by 1,471
  6. 2018AutoAugment: Learning Augmentation Policies from Datacited by 728
  7. 2017A Review on Bilevel Optimization: From Classical to Evolutionary Approaches and Applicationscited by 126
  8. 2019Data Valuation using Reinforcement Learningcited by 115
  9. 2019Adversarial AutoAugmentcited by 111
  10. 2016Medical image denoising using convolutional denoising autoencoderscited by 76
  11. 2019Deep Learning Based Detection and Correction of Cardiac MR Motion Artefacts During Reconstruction for High-Quality Segmentationcited by 17
  12. 2020Synthetic Sample Selection via Reinforcement Learningcited by 13
5 more
  1. 2019On Modelling Label Uncertainty in Deep Neural Networks: Automatic Estimation of Intra-observer Variability in 2D Echocardiography Quality Assessmentcited by 12
  2. 2020Auto-weighting for Breast Cancer Classification in Multimodal Ultrasoundcited by 9
  3. 2021Optimal MRI Undersampling Patterns for Ultimate Benefit of Medical Vision Taskscited by 7
  4. 2020Image quality assessment for closed-loop computer-assisted lung ultrasoundcited by 6
  5. 2021Learning image quality assessment by reinforcing task amenable data selectioncited by 5

Built on by

Cited by year

Citing papers per year: 2022: 1; 2023: 2; 2024: 3; 2025: 2.

8 papers.

  1. 2023Combiner and HyperCombiner Networks: Rules to Combine Multimodality MR Images for Prostate Cancer Localisationcited by 10
  2. 2022Active learning using adaptable task-based prioritisationcited by 6
  3. 2024Weakly supervised localisation of prostate cancer using reinforcement learning for bi-parametric MR imagescited by 4
  4. 2024Segmentation by registration-enabled SAM prompt engineering using five reference imagescited by 4
  5. 2023Boundary-RL: Reinforcement Learning for Weakly-Supervised Prostate Segmentation in TRUS Imagescited by 3
  6. 2025Promptable cancer segmentation using minimal expert-curated datacited by 2
  7. 2025Reasoning in machine vision by learning fast and slow thinkingcited by 1
  8. 2024Competing for pixels: a self-play algorithm for weakly-supervised segmentationcited by 0