scholarly journals Semi-Supervised Nests of Melanocytes Segmentation Method Using Convolutional Autoencoders

Sensors ◽  
2020 ◽  
Vol 20 (6) ◽  
pp. 1546 ◽  
Author(s):  
Dariusz Kucharski ◽  
Pawel Kleczek ◽  
Joanna Jaworek-Korjakowska ◽  
Grzegorz Dyduch ◽  
Marek Gorgon
Keyword(s):  
Network Architecture ◽  
State Of The Art ◽  
Ground Truth ◽  
Skin Lesions ◽  
Similarity Coefficient ◽  
Dice Similarity Coefficient ◽  
Segmentation Method ◽  
Supervised Segmentation ◽  
Malignant Skin ◽  
Histopathological Images

In this research, we present a semi-supervised segmentation solution using convolutional autoencoders to solve the problem of segmentation tasks having a small number of ground-truth images. We evaluate the proposed deep network architecture for the detection of nests of nevus cells in histopathological images of skin specimens is an important step in dermatopathology. The diagnostic criteria based on the degree of uniformity and symmetry of border irregularities are particularly vital in dermatopathology, in order to distinguish between benign and malignant skin lesions. However, to the best of our knowledge, it is the first described method to segment the nests region. The novelty of our approach is not only the area of research, but, furthermore, we address a problem with a small ground-truth dataset. We propose an effective computer-vision based deep learning tool that can perform the nests segmentation based on an autoencoder architecture with two learning steps. Experimental results verified the effectiveness of the proposed approach and its ability to segment nests areas with Dice similarity coefficient 0.81, sensitivity 0.76, and specificity 0.94, which is a state-of-the-art result.

scholarly journals Critical Aspects of Person Counting and Density Estimation

2021 ◽  
Vol 7 (2) ◽  
pp. 21
Author(s):  
Roland Perko ◽  
Manfred Klopschitz ◽  
Alexander Almer ◽  
Peter M. Roth
Keyword(s):  
Density Estimation ◽  
Network Architecture ◽  
Reference Data ◽  
State Of The Art ◽  
Limit State ◽  
Ground Truth ◽  
Data Sets ◽  
Ground Truth Generation ◽  
Baseline Approach ◽  
Critical Aspects

Many scientific studies deal with person counting and density estimation from single images. Recently, convolutional neural networks (CNNs) have been applied for these tasks. Even though often better results are reported, it is often not clear where the improvements are resulting from, and if the proposed approaches would generalize. Thus, the main goal of this paper was to identify the critical aspects of these tasks and to show how these limit state-of-the-art approaches. Based on these findings, we show how to mitigate these limitations. To this end, we implemented a CNN-based baseline approach, which we extended to deal with identified problems. These include the discovery of bias in the reference data sets, ambiguity in ground truth generation, and mismatching of evaluation metrics w.r.t. the training loss function. The experimental results show that our modifications allow for significantly outperforming the baseline in terms of the accuracy of person counts and density estimation. In this way, we get a deeper understanding of CNN-based person density estimation beyond the network architecture. Furthermore, our insights would allow to advance the field of person density estimation in general by highlighting current limitations in the evaluation protocols.

scholarly journals DEEP LEARNING BASED SKIN LESIONS DIAGNOSIS

2019 ◽  
Vol XLII-2/W12 ◽  
pp. 81-85 ◽  
Author(s):  
D. A. Gavrilov ◽  
N. N. Shchelkunov ◽  
A. V. Melerzanov
Keyword(s):  
Neural Network ◽  
Classification Accuracy ◽  
Intelligent Systems ◽  
State Of The Art ◽  
Skin Lesions ◽  
Deep Convolutional Neural Networks ◽  
Melanoma Diagnosis ◽  
Malignant Skin ◽  
Diagnosis Accuracy ◽  
Visual Diagnosis

<p><strong>Abstract.</strong> Melanoma is one of the most virulent lesions of human’s skin. The visual diagnosis accuracy of melanoma directly depends on the doctor’s qualification and specialization. State-of-the-art solutions in the field of image processing and machine learning allows to create intelligent systems based on artificial convolutional neural network exceeding human’s rates in the field of object classification, including the case of malignant skin lesions. This paper presents an algorithm for the early melanoma diagnosis based on artificial deep convolutional neural networks. The algorithm proposed allows to reach the classification accuracy of melanoma at least 91%.</p>

scholarly journals Skin Lesion Segmentation using Deep Hypercolumn Descriptors

2017 ◽  
Vol 3 (1) ◽  
Author(s):  
Dhanesh Ramachandram ◽  
Graham W. Taylor
Keyword(s):  
Image Segmentation ◽  
Skin Lesion ◽  
State Of The Art ◽  
Jaccard Index ◽  
Segmentation Method ◽  
Lesion Segmentation ◽  
Malignant Skin

We present a image segmentation method based on deep hypercolumndescriptors which produces state-of-the-art results for thesegmentation of several classes of benign and malignant skin lesions.We achieve a Jaccard index of 0.792 on the 2017 ISIC SkinLesion Segmentation Challenge dataset.

scholarly journals Rapid Automated Target Segmentation and Tracking on 4D Data without Initial Contours

2014 ◽  
Vol 2014 ◽  
pp. 1-9 ◽  
Author(s):  
Venkata V. Chebrolu ◽  
Daniel Saenz ◽  
Dinesh Tewatia ◽  
William A. Sethares ◽  
George Cannon ◽  
...  
Keyword(s):  
Software Metrics ◽  
State Of The Art ◽  
Ground Truth ◽  
Target Volume ◽  
Morphological Processing ◽  
Deformable Registration ◽  
Radiotherapy Planning ◽  
Dice Similarity Coefficient ◽  
Target Segmentation ◽  
Time Required

Purpose. To achieve rapid automated delineation of gross target volume (GTV) and to quantify changes in volume/position of the target for radiotherapy planning using four-dimensional (4D) CT.Methods and Materials. Novel morphological processing and successive localization (MPSL) algorithms were designed and implemented for achieving autosegmentation. Contours automatically generated using MPSL method were compared with contours generated using state-of-the-art deformable registration methods (usingElastix©and MIMVista software). Metrics such as the Dice similarity coefficient, sensitivity, and positive predictive value (PPV) were analyzed. The target motion tracked using the centroid of the GTV estimated using MPSL method was compared with motion tracked using deformable registration methods.Results. MPSL algorithm segmented the GTV in 4DCT images in27.0±11.1seconds per phase (512×512resolution) as compared to142.3±11.3seconds per phase for deformable registration based methods in 9 cases. Dice coefficients between MPSL generated GTV contours and manual contours (considered as ground-truth) were0.865±0.037. In comparison, the Dice coefficients between ground-truth and contours generated using deformable registration based methods were 0.909 ± 0.051.Conclusions. The MPSL method achieved similar segmentation accuracy as compared to state-of-the-art deformable registration based segmentation methods, but with significant reduction in time required for GTV segmentation.

scholarly journals Automated segmentation of the individual branches of the carotid arteries in contrast-enhanced MR angiography using DeepMedic

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Magnus Ziegler ◽  
Jesper Alfraeus ◽  
Mariana Bustamante ◽  
Elin Good ◽  
Jan Engvall ◽  
...  
Keyword(s):  
Mr Angiography ◽  
Carotid Arteries ◽  
Ground Truth ◽  
Carotid Bifurcation ◽  
External Carotid ◽  
Dice Similarity Coefficient ◽  
Segmentation Method ◽  
Positive Ratio ◽  
Contrast Enhanced ◽  
The Individual

Abstract Background Non-invasive imaging is of interest for tracking the progression of atherosclerosis in the carotid bifurcation, and segmenting this region into its constituent branch arteries is necessary for analyses. The purpose of this study was to validate and demonstrate a method for segmenting the carotid bifurcation into the common, internal, and external carotid arteries (CCA, ICA, ECA) in contrast-enhanced MR angiography (CE-MRA) data. Methods A segmentation pipeline utilizing a convolutional neural network (DeepMedic) was tailored and trained for multi-class segmentation of the carotid arteries in CE-MRA data from the Swedish CardioPulmonsary bioImage Study (SCAPIS). Segmentation quality was quantitatively assessed using the Dice similarity coefficient (DSC), Matthews Correlation Coefficient (MCC), F2, F0.5, and True Positive Ratio (TPR). Segmentations were also assessed qualitatively, by three observers using visual inspection. Finally, geometric descriptions of the carotid bifurcations were generated for each subject to demonstrate the utility of the proposed segmentation method. Results Branch-level segmentations scored DSC = 0.80 ± 0.13, MCC = 0.80 ± 0.12, F2 = 0.82 ± 0.14, F0.5 = 0.78 ± 0.13, and TPR = 0.84 ± 0.16, on average in a testing cohort of 46 carotid bifurcations. Qualitatively, 61% of segmentations were judged to be usable for analyses without adjustments in a cohort of 336 carotid bifurcations without ground-truth. Carotid artery geometry showed wide variation within the whole cohort, with CCA diameter 8.6 ± 1.1 mm, ICA 7.5 ± 1.4 mm, ECA 5.7 ± 1.0 mm and bifurcation angle 41 ± 21°. Conclusion The proposed segmentation method automatically generates branch-level segmentations of the carotid arteries that are suitable for use in further analyses and help enable large-cohort investigations.

3D Cascaded Convolutional Networks for Multi-vertebrae Segmentation

2020 ◽  
Vol 16 (3) ◽  
pp. 231-240
Author(s):  
Liu Xia ◽  
Liu Xiao ◽  
Gan Quan ◽  
Wang Bo
Keyword(s):  
Ct Images ◽  
Clinical Applications ◽  
Similarity Coefficient ◽  
Dice Similarity Coefficient ◽  
Segmentation Method ◽  
Convolutional Network ◽  
Improve Image Quality ◽  
Convolutional Networks ◽  
Spine Ct ◽  
Vertebrae Segmentation

Background: Automatic approach to vertebrae segmentation from computed tomography (CT) images is very important in clinical applications. As the intricate appearance and variable architecture of vertebrae across the population, cognate constructions in close vicinity, pathology, and the interconnection between vertebrae and ribs it is a challenge to propose a 3D automatic vertebrae CT image segmentation method. Objective: The purpose of this study was to propose an automatic multi-vertebrae segmentation method for spinal CT images. Methods: Firstly, CLAHE-Threshold-Expansion was preprocessed to improve image quality and reduce input voxel points. Then, 3D coarse segmentation fully convolutional network and cascaded finely segmentation convolutional neural network were used to complete multi-vertebrae segmentation and classification. Results: The results of this paper were compared with the other methods on the same datasets. Experimental results demonstrated that the Dice similarity coefficient (DSC) in this paper is 94.84%, higher than the V-net and 3D U-net. Conclusion: Method of this paper has certain advantages in automatically and accurately segmenting vertebrae regions of CT images. Due to the easy acquisition of spine CT images. It was proven to be more conducive to clinical application of treatment that uses our segmentation model to obtain vertebrae regions, combining with the subsequent 3D reconstruction and printing work.

scholarly journals Contrastive self-supervised clustering of scRNA-seq data

2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Madalina Ciortan ◽  
Matthieu Defrance
Keyword(s):  
Real World ◽  
Clustering Analysis ◽  
Network Architecture ◽  
Clustering Algorithm ◽  
Performance Metrics ◽  
State Of The Art ◽  
Ground Truth ◽  
Second Phase ◽  
Training Phase ◽  
Real World Datasets

Abstract Background Single-cell RNA sequencing (scRNA-seq) has emerged has a main strategy to study transcriptional activity at the cellular level. Clustering analysis is routinely performed on scRNA-seq data to explore, recognize or discover underlying cell identities. The high dimensionality of scRNA-seq data and its significant sparsity accentuated by frequent dropout events, introducing false zero count observations, make the clustering analysis computationally challenging. Even though multiple scRNA-seq clustering techniques have been proposed, there is no consensus on the best performing approach. On a parallel research track, self-supervised contrastive learning recently achieved state-of-the-art results on images clustering and, subsequently, image classification. Results We propose contrastive-sc, a new unsupervised learning method for scRNA-seq data that perform cell clustering. The method consists of two consecutive phases: first, an artificial neural network learns an embedding for each cell through a representation training phase. The embedding is then clustered in the second phase with a general clustering algorithm (i.e. KMeans or Leiden community detection). The proposed representation training phase is a new adaptation of the self-supervised contrastive learning framework, initially proposed for image processing, to scRNA-seq data. contrastive-sc has been compared with ten state-of-the-art techniques. A broad experimental study has been conducted on both simulated and real-world datasets, assessing multiple external and internal clustering performance metrics (i.e. ARI, NMI, Silhouette, Calinski scores). Our experimental analysis shows that constastive-sc compares favorably with state-of-the-art methods on both simulated and real-world datasets. Conclusion On average, our method identifies well-defined clusters in close agreement with ground truth annotations. Our method is computationally efficient, being fast to train and having a limited memory footprint. contrastive-sc maintains good performance when only a fraction of input cells is provided and is robust to changes in hyperparameters or network architecture. The decoupling between the creation of the embedding and the clustering phase allows the flexibility to choose a suitable clustering algorithm (i.e. KMeans when the number of expected clusters is known, Leiden otherwise) or to integrate the embedding with other existing techniques.

scholarly journals Accuracy and longitudinal reproducibility of quantitative femorotibial cartilage measures derived from automated U-Net-based segmentation of two different MRI contrasts: data from the osteoarthritis initiative healthy reference cohort

2020 ◽  
Author(s):  
Wolfgang Wirth ◽  
Felix Eckstein ◽  
Jana Kemnitz ◽  
Christian Frederik Baumgartner ◽  
Ender Konukoglu ◽  
...  
Keyword(s):  
Ground Truth ◽  
Dice Similarity Coefficient ◽  
Segmentation Method ◽  
Test Set ◽  
High Agreement ◽  
Set Agreement ◽  
Automated Measurements ◽  
Osteoarthritis Initiative ◽  
Reference Cohort

Abstract Objective To evaluate the agreement, accuracy, and longitudinal reproducibility of quantitative cartilage morphometry from 2D U-Net-based automated segmentations for 3T coronal fast low angle shot (corFLASH) and sagittal double echo at steady-state (sagDESS) MRI. Methods 2D U-Nets were trained using manual, quality-controlled femorotibial cartilage segmentations available for 92 Osteoarthritis Initiative healthy reference cohort participants from both corFLASH and sagDESS (n = 50/21/21 training/validation/test-set). Cartilage morphometry was computed from automated and manual segmentations for knees from the test-set. Agreement and accuracy were evaluated from baseline visits (dice similarity coefficient: DSC, correlation analysis, systematic offset). The longitudinal reproducibility was assessed from year-1 and -2 follow-up visits (root-mean-squared coefficient of variation, RMSCV%). Results Automated segmentations showed high agreement (DSC 0.89–0.92) and high correlations (r ≥ 0.92) with manual ground truth for both corFLASH and sagDESS and only small systematic offsets (≤ 10.1%). The automated measurements showed a similar test–retest reproducibility over 1 year (RMSCV% 1.0–4.5%) as manual measurements (RMSCV% 0.5–2.5%). Discussion The 2D U-Net-based automated segmentation method yielded high agreement compared with manual segmentation and also demonstrated high accuracy and longitudinal test–retest reproducibility for morphometric analysis of articular cartilage derived from it, using both corFLASH and sagDESS.

An Automatic Glioma Segmentation System Using a Multilevel Attention Pyramid Scene Parsing Network

2020 ◽  
Vol 16 ◽  
Author(s):  
Zhenyu Zhang ◽  
Shouwei Gao ◽  
Zheng Huang
Keyword(s):  
State Of The Art ◽  
Similarity Coefficient ◽  
Dice Similarity Coefficient ◽  
Training Process ◽  
Post Contrast ◽  
Image Normalization ◽  
Scene Parsing ◽  
Augmentation Techniques ◽  
Fluid Attenuated Inversion Recovery ◽  
Shape And Size

Background: Due to the significant variances in their shape and size, it is a challenging task to automatically segment gliomas. To improve the performance of glioma segmentation tasks, this paper proposed a multilevel attention pyramid scene parsing network (MLAPSPNet) that aggregates the multiscale context and multilevel features. Methods: First, T1 pre-contrast, T2-weighted fluid-attenuated inversion recovery (FLAIR) and T1 post-contrast sequences of each slice are combined to form the input. Afterward, image normalization and augmentation techniques are applied to accelerate the training process and avoid overfitting, respectively. Furthermore, the proposed MLAPSPNet that introduces multilevel pyramid pooling modules (PPMs) and attention gates is constructed. Eventually, the proposed network is compared with some existing networks. Results: The dice similarity coefficient (DSC), sensitivity and Jaccard score of the proposed system can reach 0.885, 0.933 and 0.8, respectively. The introduction of multilevel pyramid pooling modules and attention gates can improve the DSC by 0.029 and 0.022, respectively. Moreover, compared with Res-UNet, Dense-UNet, residual channel attention UNet (RCA-UNet), DeepLab V3+ and UNet++, the DSC is improved by 0.032, 0.026, 0.014, 0.041 and 0.011, respectively. Conclusion: The proposed multilevel attention pyramid scene parsing network can achieve state-of-the-art performance, and the introduction of multilevel pyramid pooling modules and attention gates can improve the performance of glioma segmentation tasks.

scholarly journals Automatic Liver Segmentation in CT Images with Enhanced GAN and Mask Region-Based CNN Architectures

2021 ◽  
Vol 2021 ◽  
pp. 1-11
Author(s):  
Xiaoqin Wei ◽  
Xiaowen Chen ◽  
Ce Lai ◽  
Yuanzhong Zhu ◽  
Hanfeng Yang ◽  
...  
Keyword(s):  
Image Segmentation ◽  
State Of The Art ◽  
Disease Diagnosis ◽  
Superior Performance ◽  
Generative Adversarial Networks ◽  
Dice Similarity Coefficient ◽  
Segmentation Method ◽  
Adversarial Networks ◽  
Liver Image

Liver image segmentation has been increasingly employed for key medical purposes, including liver functional assessment, disease diagnosis, and treatment. In this work, we introduce a liver image segmentation method based on generative adversarial networks (GANs) and mask region-based convolutional neural networks (Mask R-CNN). Firstly, since most resulting images have noisy features, we further explored the combination of Mask R-CNN and GANs in order to enhance the pixel-wise classification. Secondly, k -means clustering was used to lock the image aspect ratio, in order to get more essential anchors which can help boost the segmentation performance. Finally, we proposed a GAN Mask R-CNN algorithm which achieved superior performance in comparison with the conventional Mask R-CNN, Mask-CNN, and k -means algorithms in terms of the Dice similarity coefficient (DSC) and the MICCAI metrics. The proposed algorithm also achieved superior performance in comparison with ten state-of-the-art algorithms in terms of six Boolean indicators. We hope that our work can be effectively used to optimize the segmentation and classification of liver anomalies.

Sign in / Sign up

Export Citation Format

Share Document