Segmentation

Articles and other resources

General reviews and resources

• a github repo for medical image processing
• Methods for segmentation and classification of digital microscopy tissue images (Q. D. Vu et al.)
• A dataset and a technique for generalized nuclear segmentation for computational pathology (N. Kumar et al.)
• Deep Learning Methods for Lung Cancer Segmentation in Whole-slide Histopathology Images -- the ACDC@LungHP Challenge 2019(Z. Li et al.)
• Learning from Noisy Labels with Deep Neural Networks: A Survey (H. Song et al.)

Segmentation model architectures

• U-Net: Convolutional networks for biomedical image segmentation (O. Ronneberger, P. Fischer, T. Brox)
• Attention U-Net: Learning where to look for the pancreas (O. Oktay et al.)
• A U-Net based discriminator for generative adversarial networks (E. Schöfield, B. Schiele, A. Khoreva)
  • official implementation
• Transformers in medical imaging: a survey (F. Shamshad et al.)
• Evaluating transformer-based semantic segmentation networks for pathological image segmentation (C. Nguyen, Z. Asad, Y. Huo)
• Swin-Unet: Unet-like pure transformer for medical image segmentation (Hu Cao et al.)
  • official implementation
• Segmenter: transformer for semantic segmentation (R. Strudel et al.)
  • official implementation
• RV-GAN: segmenting retinal vascular structure in fundus photographs using a novel multi-scale generative adversarial network (S. A. Kamran et al.)
  • official implementation
• SA-UNet: spatial attention U-Net for retinal vessel segmentation (C. Guo et al.)
  • official implementation

Pretrained models

• A github repo of implemented semantic segmentation models
• U-Net for brain MRI (mateuszbuda)
• ConvNeXt for semantic segmenatiton with pretrained weights (official implementation)

Losses and metrics

• a collection of loss functions for medical segmentation
• A survey of loss functions for semantic segmentation (S. Jadon)
• Tversky loss function for image segmentation using 3D fully convolutional deep networks (S. S. M. Salehi, D. Erdogmus, A. Gholipur)
• Unified focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation (M. Yeung, E. Sala, C. Schönlieb, L. Rundo)
• On the usage of average Hausdorff distance for segmentation performance assessment: hidden error when used for ranking (O. U. Aydin et al.)
• The relationship between precision-recall and ROC curves (J. Davis, M. Goadrich)
• Boundary loss for highly unbalanced segmentation (H.Kervadec et al.)
• The Lovász-Softmax loss: A tractable surrogate for the optimization of the intersection-over-union measure in neural networks (Berman et al.)
• Segmentation of head and neck organs at risk using CNN with batch Dice loss (O. Kodym, M. Španěl, A. Herout)
  • erratum (it does not outperform the reference method when evaluated correctly)
• Semantic segmentation with labeling uncertainty and class imbalance (P. O. Bressan et al.)
• Are we using appropriate segmentation metrics? Identifying correlates of human expert perception for CNN training beyond rolling the DICE coefficient (F. Kofler et al.)

Methods for dealing with imbalanced data

• Enhancing the reliability of out-of-distribution image detection in neural networks (S. Liang, Y. Li, R. Srikant)
• An insight into classification with imbalanced data: Empirical results and current trends on using data intrinsic characteristics (V. López et al.)
• Domain adaptation for semantic segmentation via class-balanced self-training (Y. Zou, Z. Yu, B. V. K. V. Kumar, J. Wang)
• Handling unbalanced data in deep image segmentation (H. Small, J. Ventura)
• SMOTE: Synthetic Minority Over-sampling Technique (N. V. Chawla, K. W. Bowyer, L. O. Hall, W. P. Kegelmeyer)
  • this focuses on tabular data; for computer vision, image augmentation techniques are probably better
• a python package for imbalanced data

Methods for dealing with noisy data

• Embracing imperfect datasets: A review of deep learning solutions for medical image segmentation (N. Tajbakhsh et al.)
• Learning confidence for out-of-distribution detection in neural networks (T. DeVries, G. W. Taylor)

Histpathology datasets

• Lizard: A large-scale dataset for colonic nuclear instance segmentation and classification (S. Graham et al.)
• Multiple instance captioning: Learning representations from histopathology textbooks and articles (J. Gamper, N. Rajpoot)

Performance enhancing hacks

• Adaptive thresholding: A comparative study (N. Dey, G. Dey, S. Dutta, S. Chakraborty)
• Adaptive threshold segmentation of pituitary adenomas from FDG PET images for radiosurgery (H. M. Thomas et al.)