Persistent Homology for Fast Tumor Segmentation in Whole Slide Histology Images

Qaiser, Talha; Sirinukunwattana, Korsuk; Nakane, Kazuaki; Tsang, Yee-Wah; Epstein, D. B. A.; Rajpoot, Nasir M.

doi:10.1016/j.procs.2016.07.033

Cited by 56 publications

(43 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, we split the IQR of each class into Q same size bins and select the value that lies closest to the median of corresponding bin, where Q represents the number of exemplar patches for each class. This method for selection of exemplar patches differs from Qaiser et al (2016) and is capable of handling a much larger dataset as in Section 5.2.1.…”

Section: Selection Of Exemplar Patchesmentioning

confidence: 99%

“…In total, we extracted, 75,000 patches for training from 50 WSIs and 37,500 patches for testing from 25 WSIs. The collected dataset for this study is roughly 20 times more than that in Qaiser et al (2016) and at least 2 times more than in Qaiser et al (2017). For generating the tumor probability map of a WSI, we first split the given WSI into patches and then applied our methods to each patch.…”

Section: Dataset and Experimental Setupmentioning

confidence: 99%

“…Then we selected those patches that lie closest to the cluster centroids as exemplar patches, one patch per centroid. The third algorithm for comparison is to select highly activated patches as exemplars from a CNN as proposed in Qaiser et al (2016). For a fair comparison, an equal number of exemplar patches were selected for each of the above mentioned algorithms.…”

Section: Selection Of Exemplar Patchesmentioning

confidence: 99%

See 2 more Smart Citations

Fast and accurate tumor segmentation of histology images using persistent homology and deep convolutional features

Qaiser

Tsang

Taniyama

et al. 2019

Medical Image Analysis

Self Cite

160

View full text Add to dashboard Cite

Tumor segmentation in whole-slide images of histology slides is an important step towards computer-assisted diagnosis. In this work, we propose a tumor segmentation framework based on the novel concept of persistent homology profiles (PHPs). For a given image patch, the homology profiles are derived by efficient computation of persistent homology, which is an algebraic tool from homology theory. We propose an efficient way of computing topological persistence of an image, alternative to simplicial homology. The PHPs are devised to distinguish tumor regions from their normal counterparts by modeling the atypical characteristics of tumor nuclei. We propose two variants of our method for tumor segmentation: one that targets speed without compromising accuracy and the other that targets higher accuracy. The fast version is based on selection of exemplar image patches from a convolution neural network (CNN) and patch classification by quantifying the divergence between the PHPs of exemplars and the input image patch. Detailed comparative evaluation shows that the proposed algorithm is significantly faster than competing algorithms while achieving comparable results. The accurate version combines the PHPs and high-level CNN features and employs a multi-stage ensemble strategy for image patch labeling. Experimental results demonstrate that the combination of 1 arXiv:1805.03699v1 [cs.CV] 9 May 2018 PHPs and CNN features outperforms competing algorithms. This study is performed on two independently collected colorectal datasets containing adenoma, adenocarcinoma, signet and healthy cases. Collectively, the accurate tumor segmentation produces the highest average patch-level F1-score, as compared with competing algorithms, on malignant and healthy cases from both the datasets.Overall the proposed framework highlights the utility of persistent homology for histopathology image analysis.

show abstract

Section: Selection Of Exemplar Patchesmentioning

confidence: 99%

Section: Dataset and Experimental Setupmentioning

confidence: 99%

Section: Selection Of Exemplar Patchesmentioning

confidence: 99%

See 1 more Smart Citation

Fast and accurate tumor segmentation of histology images using persistent homology and deep convolutional features

Qaiser

Tsang

Taniyama

et al. 2019

Medical Image Analysis

Self Cite

160

View full text Add to dashboard Cite

show abstract

“…Our approach aims to fit the desired absence or presence of certain features and so complex features can be penalised or rewarded, as is appropriate for the task at hand. Persistent homology has previously been applied to the problem of semantic segmentation, such as in [9,16,2]. The important distinction between our method and these previous works is that they apply PH to the input image to extract features, which are then used as inputs to some other algorithm for training.…”

Section: Related Workmentioning

confidence: 99%

Explicit Topological Priors for Deep-Learning Based Image Segmentation Using Persistent Homology

Clough

Öksüz

Byrne

et al. 2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

We present a novel method to explicitly incorporate topological prior knowledge into deep learning based segmentation, which is, to our knowledge, the first work to do so. Our method uses the concept of persistent homology, a tool from topological data analysis, to capture high-level topological characteristics of segmentation results in a way which is differentiable with respect to the pixelwise probability of being assigned to a given class. The topological prior knowledge consists of the sequence of desired Betti numbers of the segmentation. As a proof-of-concept we demonstrate our approach by applying it to the problem of left-ventricle segmentation of cardiac MR images of 500 subjects from the UK Biobank dataset, where we show that it improves segmentation performance in terms of topological correctness without sacrificing pixelwise accuracy.

show abstract

“…Convolutional neural networks (CNNs) have been widely used to achieve the state-of-the-art results for different histology image analysis tasks such as nuclei detection and classification [2], [3], [4], metastasis detection [5], [6], [7], tumor segmentation [8] and cancer grading [9], [10], [11]. Each task requires a different amount of contextual information, for instance, cell classification needs only high-resolution cell appearance along with little neighboring tissue whereas tumour detection and segmentation rely on a larger context covering multiple cells simultaneously.…”

mentioning

confidence: 99%

Context-Aware Convolutional Neural Network for Grading of Colorectal Cancer Histology Images

Shaban

Awan

Fraz

et al. 2020

IEEE Trans. Med. Imaging

Self Cite

127

View full text Add to dashboard Cite

Digital histology images are amenable to the application of convolutional neural network (CNN) for analysis due to the sheer size of pixel data present in them. CNNs are generally used for representation learning from small image patches (e.g. 224 × 224) extracted from digital histology images due to computational and memory constraints. However, this approach does not incorporate high-resolution contextual information in histology images. We propose a novel way to incorporate larger context by a context-aware neural network based on images with a dimension of 1, 792 × 1, 792 pixels. The proposed framework first encodes the local representation of a histology image into high dimensional features then aggregates the features by considering their spatial organization to make a final prediction. The proposed method is evaluated for colorectal cancer grading and breast cancer classification. A comprehensive analysis of some variants of the proposed method is presented. Our method outperformed the traditional patch-based approaches, problem-specific methods, and existing context-based methods quantitatively by a margin of 3.61%. Code and dataset related information is available at this link: https://tia-lab.github.io/Context-Aware-CNN

show abstract

Persistent Homology for Fast Tumor Segmentation in Whole Slide Histology Images

Cited by 56 publications

References 7 publications

Fast and accurate tumor segmentation of histology images using persistent homology and deep convolutional features

Fast and accurate tumor segmentation of histology images using persistent homology and deep convolutional features

Explicit Topological Priors for Deep-Learning Based Image Segmentation Using Persistent Homology

Context-Aware Convolutional Neural Network for Grading of Colorectal Cancer Histology Images

Contact Info

Product

Resources

About