Automated Segmentation of DCIS in Whole Slide Images

Seth, Nikhil; Akbar, Shazia; Nofech‐Mozes, Sharon; Salama, Sherine; Martel, Anne L.

doi:10.1007/978-3-030-23937-4_8

Cited by 8 publications

(7 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The whole FCN model can be trained via end-to-end backpropagation and directly outputs a dense per-pixel prediction score map. Hence, segmentation models in histopathology are mainly built on the representative power of FCN and its variants, which are generally formulated as a semantic segmentation task, with applications ranging from nucleus/gland/duct segmentation (Kumar et al, 2019;Sirinukunwattana et al, 2017;Seth et al, 2019) to the prediction of cancer (Liu et al, 2019;Bulten et al, 2019a) in WSIs.…”

Section: Segmentation Modelsmentioning

confidence: 99%

“…A considerable amount of work has been done to combine low and highresolution inputs in making better decisions in various forms and problems (Li et al, 2019b;Chang et al, 2017;Shujun Wang et al, 2019;. However, it is still unclear that these methods are more effective in segmentation tasks compared to selecting a single "best fit" resolution (Seth et al, 2019).…”

Section: Challenges In Histopathology Image Analysismentioning

confidence: 99%

“…Obtaining a well-curated data set is, however, often expensive and requires significant manual expertise to obtain clean and accurate annotations. There will always be variability between pathologists so ideally the inter-observer agreement should be quantified (Bulten et al, 2019b;Akbar et al, 2019;Seth et al, 2019) and if possible, a consensus between pathologists reached (Veta et al, 2015). Some attempts have been made to generate additional annotated data by using alternative techniques like data augmentation (Tellez et al, 2019a), image synthesis (Hou et al, 2019a) and crowdsourcing (Albarqouni et al, 2016), but it is not yet clear that they are appropriate for digital pathology.…”

Section: Quality Of Training and Validation Datamentioning

confidence: 99%

See 2 more Smart Citations

Deep neural network models for computational histopathology: A survey

Srinidhi,

Ciga,

Martel

2019

Preprint

Self Cite

View full text Add to dashboard Cite

Histopathological images contain rich phenotypic information that can be used to monitor underlying mechanisms contributing to diseases progression and patient survival outcomes. Recently, deep learning has become the mainstream methodological choice for analyzing and interpreting cancer histology images. In this paper, we present a comprehensive review of state-of-the-art deep learning approaches that have been used in the context of histopathological image analysis. From the survey of over 130 papers, we review the field's progress based on the methodological aspect of different machine learning strategies such as supervised, weakly supervised, unsupervised, transfer learning and various other sub-variants of these methods. We also provide an overview of deep learning based survival models that are applicable for disease-specific prognosis tasks. Finally, we summarize several existing open datasets and highlight critical challenges and limitations with current deep learning approaches, along with possible avenues for future research.

show abstract

Section: Segmentation Modelsmentioning

confidence: 99%

Section: Challenges In Histopathology Image Analysismentioning

confidence: 99%

Section: Quality Of Training and Validation Datamentioning

confidence: 99%

See 1 more Smart Citation

Deep neural network models for computational histopathology: A survey

Srinidhi,

Ciga,

Martel

2019

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…To exploit all the available GPU memory thus requires one to trade-off the field of view (FoV) i.e., spatial extent of context, against the spatial resolutions i.e., level of image detail. Tuning this trade-off exhaustively is expensive [6], and as a result, it is commonly set based on crude developers' intuitions. Moreover, as [6] points out, the optimal trade-off is application or even image dependent (e.g., some parts of the images may require more context than local details), and thus the existence of an "one-size-fits-all" FoV/resolution trade-off for a given application is highly questionable.…”

Section: Introductionmentioning

confidence: 99%

Foveation for Segmentation of Ultra-High Resolution Images

Jin,

Tanno,

et al. 2020

Preprint

View full text Add to dashboard Cite

Segmentation of ultra-high resolution images is challenging because of their enormous size, consisting of millions or even billions of pixels. Typical solutions include dividing input images into patches of fixed size and/or down-sampling to meet memory constraints. Such operations incur information loss in the field-of-view (FoV) i.e., spatial coverage and the image resolution. The impact on segmentation performance is, however, as yet understudied. In this work, we start with a motivational experiment which demonstrates that the trade-off between FoV and resolution affects the segmentation performance on ultra-high resolution images-and furthermore, its influence also varies spatially according to the local patterns in different areas. We then introduce foveation module, a learnable "dataloader" which, for a given ultra-high resolution image, adaptively chooses the appropriate configuration (FoV/resolution trade-off) of the input patch to feed to the downstream segmentation model at each spatial location of the image. The foveation module is jointly trained with the segmentation network to maximise the task performance. We demonstrate on three publicly available high-resolution image datasets that the foveation module consistently improves segmentation performance over the cases trained with patches of fixed FoV/resolution trade-off. Our approach achieves the SoTA performance on the DeepGlobe aerial image dataset. On the Gleason2019 histopathology dataset, our model achieves better segmentation accuracy for the two most clinically important and ambiguous classes (Gleason Grade 3 and 4) than the top performers in the challenge by 13.1% and 7.5%, and improves on the average performance of 6 human experts by 6.5% and 7.5%. Our code and trained models are available at https://github.com/lxasqjc/Foveation-Segmentation.

show abstract

“…level of image detail. Tuning this trade-off exhaustively is expensive [2], and as a result, it is commonly set by crude developer intuition.…”

Section: Introductionmentioning

confidence: 99%

Foveation for Segmentation of Mega-Pixel Histology Images

Chen

Tanno

et al. 2020

Medical Image Computing and Computer Assisted Intervention – MICCAI 2020

View full text Add to dashboard Cite

Segmenting histology images is challenging because of the sheer size of the images with millions or even billions of pixels. Typical solutions pre-process each histology image by dividing it into patches of fixed size and/or down-sampling to meet memory constraints. Such operations incur information loss in the field-of-view (FoV) (i.e., spatial coverage) and the image resolution. The impact on segmentation performance is, however, as yet understudied. In this work, we first show under typical memory constraints (e.g., 10G GPU memory) that the trade-off between FoV and resolution considerably affects segmentation performance on histology images, and its influence also varies spatially according to local patterns in different areas (see Fig. 1). Based on this insight, we then introduce foveation module, a learnable "dataloader" which, for a given histology image, adaptively chooses the appropriate configuration (FoV/resolution trade-off) of the input patch to feed to the downstream segmentation model at each spatial location (Fig. 1). The foveation module is jointly trained with the segmentation network to maximise the task performance. We demonstrate, on the Gleason2019 challenge dataset for histopathology segmentation, that the foveation module improves segmentation performance over the cases trained with patches of fixed FoV/resolution trade-off. Moreover, our model achieves better segmentation accuracy for the two most clinically important and ambiguous classes (Gleason Grade 3 and 4) than the top performers in the challenge by 13.1% and 7.5%, and improves on the average performance of 6 human experts by 6.5% and 7.5%.

show abstract

Automated Segmentation of DCIS in Whole Slide Images

Cited by 8 publications

References 3 publications

Deep neural network models for computational histopathology: A survey

Deep neural network models for computational histopathology: A survey

Foveation for Segmentation of Ultra-High Resolution Images

Foveation for Segmentation of Mega-Pixel Histology Images

Contact Info

Product

Resources

About