Physically Nonlinear Bending of Composite Plates

Yakubovskiy, Yuriy; Kolosov, V I

doi:10.1007/978-3-030-42351-3_27

Cited by 6 publications

(8 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We carried out two experiments using three different acquisition functions for iterative learning on the task of on the task of retinal layer segmentation with U-Net and EfficientNet as the network backbone[15]. The three acquisition functions are (i) random selection (no active learning), (ii) a SOTA active learning method using epistemic uncertainty only[7, 8], and (iii) active learning with similarity and epistemic uncertainty (ours).…”

Section: Methodsmentioning

confidence: 99%

Deep Active Learning for Robust Biomedical Segmentation

Arikan¹,

Sallo

Montesel

et al. 2023

Preprint

View full text Add to dashboard Cite

Deep learning for medical applications faces many unique challenges. A major challenge is the large amount of labelled data for training, while working in a relatively data scarce environment. Active learning can be used to overcome the vast data need challenge. A second challenged faced is poor performance outside of a experimental setting, contrary to the high requirement for safety and robustness. In this paper, we present a novel framework for estimating uncertainty metrics and incorporating a similarity measure to improve active learning strategies. To showcase effectiveness, a medical image segmentation task was used as an exemplar. In addition to faster learning, robustness was also addressed through adversarial perturbations. Using epistemic uncertainty and our framework, we can cut number of annotations needed by 39% and by 54% using epistemic uncertainty and a similarity metric.

show abstract

Section: Methodsmentioning

confidence: 99%

Deep Active Learning for Robust Biomedical Segmentation

Arikan¹,

Sallo

Montesel

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Specifically, the Keras ImageDataGenerator [ 47 ] was used to create augmentation generators for the training and validation data. Publicly available pretrained EfficientNet models were utilized [ 48 ].…”

Section: Methods and Techniquesmentioning

confidence: 99%

Classification of Hematoxylin and Eosin‐Stained Breast Cancer Histology Microscopy Images Using Transfer Learning with EfficientNets

Munien

Viriri

2021

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Breast cancer is a fatal disease and is a leading cause of death in women worldwide. The process of diagnosis based on biopsy tissue is nontrivial, time-consuming, and prone to human error, and there may be conflict about the final diagnosis due to interobserver variability. Computer-aided diagnosis systems have been designed and implemented to combat these issues. These systems contribute significantly to increasing the efficiency and accuracy and reducing the cost of diagnosis. Moreover, these systems must perform better so that their determined diagnosis can be more reliable. This research investigates the application of the EfficientNet architecture for the classification of hematoxylin and eosin-stained breast cancer histology images provided by the ICIAR2018 dataset. Specifically, seven EfficientNets were fine-tuned and evaluated on their ability to classify images into four classes: normal, benign, in situ carcinoma, and invasive carcinoma. Moreover, two standard stain normalization techniques, Reinhard and Macenko, were observed to measure the impact of stain normalization on performance. The outcome of this approach reveals that the EfficientNet-B2 model yielded an accuracy and sensitivity of 98.33% using Reinhard stain normalization method on the training images and an accuracy and sensitivity of 96.67% using the Macenko stain normalization method. These satisfactory results indicate that transferring generic features from natural images to medical images through fine-tuning on EfficientNets can achieve satisfactory results.

show abstract

“…We used a CNN of the U-Net [26, 27] family with an input size of 512 × 512 pixels and four encoder stages. We employed transfer learning: the original encoders were replaced by EfficientNet-B3 [28] encoders that were pretrained on the ImageNet dataset [29].…”

Section: Methodsmentioning

confidence: 99%

A representative reference for MRI-based human axon radius assessment using light microscopy

Mordhorst

Morozova

Papazoglou

et al. 2021

Preprint

View full text Add to dashboard Cite

Non-invasive assessment of axon radii via MRI bears great potential for clinical and neuroscience research as it is a main determinant of the neuronal conduction velocity. However, there is a lack of representative histological reference data on the scale of the cross-section of MRI voxels for validating the MRI-visible, effective radius (reff). Because the current gold standard stems from neuroanatomical studies designed to estimate the frequency-weighted arithmetic mean radius (rarith) on small ensembles of axons, it is unsuited to estimate the tail-weighted reff. We propose CNN-based segmentation on high-resolution, large-scale light microscopy (lsLM) data to generate a representative reference for reff. In a human corpus callosum, we assessed estimation accuracy and bias of rarith and reff. Furthermore, we investigated whether mapping anatomy-related variation of rarith and reff is confounded by low-frequency variation of the image intensity, e.g., due to staining heterogeneity. Finally, we analyzed the potential error due to outstandingly large axons in reff. Compared to rarith, reff was estimated with higher accuracy (normalized-root-mean-square-error of reff: 7.2 %; rarith: 21.5 %) and lower bias (normalized-mean-bias-error of reff: -1.7 %; rarith: 16 %). While rarith was confounded by variation of the image intensity, variation of reff seemed anatomy-related. The largest axons contributed between 0.9 % and 3 % to reff. In conclusion, the proposed method accurately estimates reff at MRI voxel resolution across a human corpus callosum sample. Further investigations are required to assess generalization to brain areas with different axon radii ensembles.

show abstract

Physically Nonlinear Bending of Composite Plates

Cited by 6 publications

References 15 publications

Deep Active Learning for Robust Biomedical Segmentation

Deep Active Learning for Robust Biomedical Segmentation

Classification of Hematoxylin and Eosin‐Stained Breast Cancer Histology Microscopy Images Using Transfer Learning with EfficientNets

A representative reference for MRI-based human axon radius assessment using light microscopy

Contact Info

Product

Resources

About