Blob Loss: Instance Imbalance Aware Loss Functions for Semantic Segmentation

@article{Kofler2022BlobLI,
  title={Blob Loss: Instance Imbalance Aware Loss Functions for Semantic Segmentation},
  author={Florian Kofler and Suprosanna Shit and Ivan Ezhov and L Fidon and Rami Al-Maskari and Hongwei Li and Harsharan Bhatia and Timo Loehr and Marie Piraud and Ali Erturk and Jan Stefan Kirschke and Jan C. Peeken and Tom Kamiel Magda Vercauteren and Claus Zimmer and Benedikt Wiestler and Bjoern H Menze},
  journal={ArXiv},
  year={2022},
  volume={abs/2205.08209}
}
. Deep convolutional neural networks (CNN) have proven to be remarkably effective in semantic segmentation tasks. Most popular loss functions were introduced targeting improved volumetric scores, such as the Sørensen–Dice coefficient (DSC). By design, DSC score and sensitivity . Blob loss is designed for semantic segmentation problems in which the instances are the connected components within a class. We extensively evaluate a DSC-based blob loss in five complex 3D semantic segmentation tasks… 

Where is VALDO? VAscular Lesions Detection and segmentatiOn challenge at MICCAI 2021

Results showed a large variability in performance both across teams and across tasks, with promising results notably for Task 1 - EPVS and Task 2 - Microbleeds and not practically useful results yet for Task 3 - Lacunes.

Metrics reloaded: Pitfalls and recommendations for image analysis validation

The Metrics Reloaded framework was developed in a multi-stage Delphi process and is based on the novel concept of a problem fingerprint – a structured representation of the given problem that captures all aspects that are relevant for metric selection from the domain interest to the properties of the target structure(s), data set and algorithm output.

Are we using appropriate segmentation metrics? Identifying correlates of human expert perception for CNN training beyond rolling the DICE coefficient

It is discovered that current standard metrics and loss functions correlate only moderately with the segmentation quality assessment of experts, and a novel strat-egy employing techniques of classical statistics to create complementary compound loss functions to better approximate human expert perception for convolutional neural network (CNN) training is proposed.

References

SHOWING 1-10 OF 32 REFERENCES

Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentations

This work investigates the behavior of these loss functions and their sensitivity to learning rate tuning in the presence of different rates of label imbalance across 2D and 3D segmentation tasks and proposes to use the class re-balancing properties of the Generalized Dice overlap as a robust and accurate deep-learning loss function for unbalanced tasks.

Optimizing Intersection-Over-Union in Deep Neural Networks for Image Segmentation

This paper proposes an approach for directly optimizing this intersection-over-union (IoU) measure in deep neural networks and demonstrates that this approach outperforms DNNs trained with standard softmax loss.

Tversky Loss Function for Image Segmentation Using 3D Fully Convolutional Deep Networks

A generalized loss function based on the Tversky index is proposed to address the issue of data imbalance and achieve much better trade-off between precision and recall in training 3D fully convolutional deep neural networks.

Loss odyssey in medical image segmentation

V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation

This work proposes an approach to 3D image segmentation based on a volumetric, fully convolutional, neural network, trained end-to-end on MRI volumes depicting prostate, and learns to predict segmentation for the whole volume at once.

Universal Loss Reweighting to Balance Lesion Size Inequality in 3D Medical Image Segmentation

A loss reweighting approach to increase the ability of the network to detect small lesions and shows that inverse weighting considerably increases the detection quality, while preserves the delineation quality on a state-of-the-art level.

Are we using appropriate segmentation metrics? Identifying correlates of human expert perception for CNN training beyond rolling the DICE coefficient

It is discovered that current standard metrics and loss functions correlate only moderately with the segmentation quality assessment of experts, and a novel strat-egy employing techniques of classical statistics to create complementary compound loss functions to better approximate human expert perception for convolutional neural network (CNN) training is proposed.

Optimization for Medical Image Segmentation: Theory and Practice When Evaluating With Dice Score or Jaccard Index

It is confirmed that metric-sensitive losses are superior to cross-entropy based loss functions in case of evaluation with Dice Score or Jaccard Index in a multi-class setting, and across different object sizes and foreground/background ratios.

U-Net: Convolutional Networks for Biomedical Image Segmentation

It is shown that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks.

Nucleus segmentation across imaging experiments: the 2018 Data Science Bowl

The 2018 Data Science Bowl attracted 3,891 teams worldwide to make the first attempt to build a segmentation method that could be applied to any two-dimensional light microscopy image of stained nuclei across experiments, with no human interaction.