Tissue Type Recognition in Whole Slide Histological Images

Khvostikov, Alexander; Krylov, Andrey S.; Михайлов, И. А.; Malkov, P.G.; Данилова, Н В

doi:10.20948/graphicon-2021-3027-496-507

Cited by 10 publications

(2 citation statements)

References 10 publications

(11 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The resulting model still outperforms all other solutions by over 3% while being insensitive to variations in staining, noise, sharpness and lighting conditions, thus significantly reducing the impact of any potential batch effect present in the data. Method BA, % Accuracy, % DenseNet based solution [42] 90.3 92.9 VGG19 based solution [40] 94.3 Inception-v3 based solution [86] 94.8 ResNet-50 based solution [75] 94.8 VGG16 based solution [4] 95.3 CONCH (ViT-Base foundation transformer model) [50] 93.0 -iBOT (ViT-Large transformer model) [19] 94.4 95.8 DINO (ViT transformer model) [38] 94.5 95.9 Ensemble of 4 models (DenseNet, IncResNetV2, Xception and custom) [24] 96.16 EfficientNet-B7 (ImageNet initialized weights) 94.76 96.18 Ensemble of 5 models (Same as [24] + VGG16) [46] 96.26 CTransPath [88] 96.52 DeepCMorph 95.59 96.99…”

Section: Pan Cancer Tcga Data Classificationmentioning

confidence: 99%

Histopathological image classification with deep convolutional neural networks

Alom

Aspiras

Taha

et al. 2019

Applications of Machine Learning

View full text Add to dashboard Cite

Histopathological images are widely used for the analysis of diseased (tumor) tissues and patient treatment selection. While the majority of microscopy image processing was previously done manually by pathologists, recent advances in computer vision allow for accurate recognition of lesion regions with deep learning-based solutions. Such models, however, usually require extensive annotated datasets for training, which is often not the case in the considered task, where the number of available patient data samples is very limited. To deal with this problem, we propose a novel DeepCMorph model pre-trained to learn cell morphology and identify a large number of different cancer types. The model consists of two modules: the first one performs cell nuclei segmentation and annotates each cell type, and is trained on a combination of 8 publicly available datasets to ensure its high generalizability and robustness. The second module combines the obtained segmentation map with the original microscopy image and is trained for the downstream task. We pre-trained this module on the Pan-Cancer TCGA dataset consisting of over 270K tissue patches extracted from 8736 diagnostic slides from 7175 patients. The proposed solution achieved a new state-ofthe-art performance on the dataset under consideration, detecting 32 cancer types with over 82% accuracy and outperforming all previously proposed solutions by more than 4%. We demonstrate that the resulting pre-trained model can be easily fine-tuned on smaller microscopy datasets, yielding superior results compared to the current top solutions and models initialized with ImageNet weights. The codes and pre-trained models presented in this paper are available at: https://github.com/aiff22/DeepCMorph

show abstract

Section: Pan Cancer Tcga Data Classificationmentioning

confidence: 99%

Histopathological image classification with deep convolutional neural networks

Alom

Aspiras

Taha

et al. 2019

Applications of Machine Learning

View full text Add to dashboard Cite

show abstract

“…Currently, some medical imaging tasks, such as histological image classification, are solved using neural networks [9], which are vulnerable to adversarial attacks. Hence, making predictions using neural networks in medical imaging can be dangerous.…”

Section: Introductionmentioning

confidence: 99%

Augmenting Histological Images with Adversarial Attacks

Lockshin

Khvostikov

Krylov

2022

Proceedings of the 32nd International Conference on Computer Graphics and Vision

Self Cite

View full text Add to dashboard Cite

Neural networks have shown to be vulnerable against adversarial attacks - images with carefully crafted adversarial noise that is imperceptible to the human eye. In medical imaging tasks this can be a major threat for making predictions based on deep neural network solutions. In this paper we propose a pipeline for augmenting a small histological image dataset using State-of-the-Art data generation methods and demonstrate an increase in accuracy of a neural classifier trained on the augmented dataset when faced with adversarial images. When trained on the non-augmented dataset, the neural network achieves an accuracy of 55.24 on the test set with added adversarial noise, and an accuracy of 97.40 on the same test set when trained on the augmented dataset.

show abstract

TAHIR: Transformer-Based Affine Histological Image Registration

Pyatov

Sorokin

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Tissue Type Recognition in Whole Slide Histological Images

Cited by 10 publications

References 10 publications

Histopathological image classification with deep convolutional neural networks

Histopathological image classification with deep convolutional neural networks

Augmenting Histological Images with Adversarial Attacks

TAHIR: Transformer-Based Affine Histological Image Registration

Contact Info

Product

Resources

About