CNN Segmentation Research Articles

Automated intracranial vessel segmentation of 4D flow MRI data in patients with atherosclerotic stenosis using a convolutional neural network.

Intracranial 4D flow MRI enables quantitative assessment of hemodynamics in patients with intracranial atherosclerotic disease (ICAD). However, quantitative assessments are still challenging due to the time-consuming vessel segmentation, especially in the presence of stenoses, which can often result in user variability. To improve the reproducibility and robustness as well as to accelerate data analysis, we developed an accurate, fully automated segmentation for stenosed intracranial vessels using deep learning. 154 dual-VENC 4D flow MRI scans (68 ICAD patients with stenosis, 86 healthy controls) were retrospectively selected. Manual segmentations were used as ground truth for training. For automated segmentation, deep learning was performed using a 3D U-Net. 20 randomly selected cases (10 controls, 10 patients) were separated and solely used for testing. Cross-sectional areas and flow parameters were determined in the Circle of Willis (CoW) and the sinuses. Furthermore, the flow conservation error was calculated. For statistical comparisons, Dice scores (DS), Hausdorff distance (HD), average symmetrical surface distance (ASSD), Bland-Altman analyses, and interclass correlations were computed using the manual segmentations from two independent observers as reference. Finally, three stenosis cases were analyzed in more detail by comparing the 4D flow-based segmentations with segmentations from black blood vessel wall imaging (VWI). Training of the network took approximately 10 h and the average automated segmentation time was 2.2 ± 1.0 s. No significant differences in segmentation performance relative to two independent observers were observed. For the controls, mean DS was 0.85 ± 0.03 for the CoW and 0.86 ± 0.06 for the sinuses. Mean HD was 7.2 ± 1.5 mm (CoW) and 6.6 ± 3.7 mm (sinuses). Mean ASSD was 0.15 ± 0.04 mm (CoW) and 0.22 ± 0.17 mm (sinuses). For the patients, the mean DS was 0.85 ± 0.04 (CoW) and 0.82 ± 0.07 (sinuses), the HD was 8.4 ± 3.1 mm (CoW) and 5.7 ± 1.9 mm (sinuses) and the mean ASSD was 0.22 ± 0.10 mm (CoW) and 0.22 ± 0.11 mm (sinuses). Small bias and limits of agreement were observed in both cohorts for the flow parameters. The assessment of the cross-sectional lumen areas in stenosed vessels revealed very good agreement (ICC: 0.93) with the VWI segmentation but a consistent overestimation (bias ± LOA: 28.1 ± 13.9%). Deep learning was successfully applied for fully automated segmentation of stenosed intracranial vasculatures using 4D flow MRI data. The statistical analysis of segmentation and flow metrics demonstrated very good agreement between the CNN and manual segmentation and good performance in stenosed vessels. To further improve the performance and generalization, more ICAD segmentations as well as other intracranial vascular pathologies will be considered in the future.

3118: CNN vs manual segmentation for ES-NSCLC: performance assessment and current pitfalls

3118: CNN vs manual segmentation for ES-NSCLC: performance assessment and current pitfalls

Automatic Finding of Brain-Tumour Group Using CNN Segmentation and Moth-Flame-Algorithm, Selected Deep and Handcrafted Features

Automatic Finding of Brain-Tumour Group Using CNN Segmentation and Moth-Flame-Algorithm, Selected Deep and Handcrafted Features

Hybrid Algorithm for the Detection of Lung Cancer Using CNN and Image Segmentation

When it comes to cancer-related fatalities, lung cancer is by far the most common cause. Early detection is the key to a successful diagnosis and treatment plan for lung cancer, just as it is for other types of cancer. Automatic CAD systems for lung cancer screening using computed tomography scans primarily involve two steps: the first step is to detect all potentially malignant pulmonary nodules, and the second step is to determine whether or not the nodules are malignant. There have been a lot of books published recently about the first phase, but not many about the second stage. Screening for lung cancer requires a careful investigation on each suspicious nodule and the integration of information from all nodules. This is because the presence of pulmonary nodules does not always indicate cancer, and the morphology of nodules, including their shape, size, and contextual information, has a complex relationship with cancer. In order to overcome this problem, we suggest a deep CNN architecture that is different from the architectures that are commonly utilised in computer vision. After the suspicious nodules have been formed with the modified version of U-Net, they are used as an input data for our model. First, the suspicious nodules are generated with U-Net. To automatically diagnose lung cancer, the suggested model is a multi-path CNN that concurrently makes use of local characteristics as well as more general contextual characteristics from a wider geographical area. In order to accomplish this, the model consisted of three separate pathways, each of which used a different receptive field size, which contributed to the modelling of distant dependencies (short and long-range dependencies of the neighbouring pixels). After that, we concatenate characteristics from the three different pathways in order to further improve our model's performance. In conclusion, one of the contributions that we have made is the development of a retraining phase system. This system enables us to address issues that are caused by an uneven distribution of picture labels. The experimental findings from the KDSB 2017 challenge demonstrate that our model is more adaptable to the described inconsistency among the nodules' sizes and shapes. Furthermore, our model obtained better results in comparison to other researches.

Tea leaf disease detection and identification based on YOLOv7 (YOLO-T)

A reliable and accurate diagnosis and identification system is required to prevent and manage tea leaf diseases. Tea leaf diseases are detected manually, increasing time and affecting yield quality and productivity. This study aims to present an artificial intelligence-based solution to the problem of tea leaf disease detection by training the fastest single-stage object detection model, YOLOv7, on the diseased tea leaf dataset collected from four prominent tea gardens in Bangladesh. 4000 digital images of five types of leaf diseases are collected from these tea gardens, generating a manually annotated, data-augmented leaf disease image dataset. This study incorporates data augmentation approaches to solve the issue of insufficient sample sizes. The detection and identification results for the YOLOv7 approach are validated by prominent statistical metrics like detection accuracy, precision, recall, mAP value, and F1-score, which resulted in 97.3%, 96.7%, 96.4%, 98.2%, and 0.965, respectively. Experimental results demonstrate that YOLOv7 for tea leaf diseases in natural scene images is superior to existing target detection and identification networks, including CNN, Deep CNN, DNN, AX-Retina Net, improved DCNN, YOLOv5, and Multi-objective image segmentation. Hence, this study is expected to minimize the workload of entomologists and aid in the rapid identification and detection of tea leaf diseases, thus minimizing economic losses.

Two-Stage Cascaded CNN Model for 3D Mitochondria EM Segmentation

Mitochondria are the organelles that generate energy for the cells. Many studies have suggested that mitochondrial dysfunction or impairment may be related to cancer and other neurodegenerative disorders such as Alzheimer’s and Parkinson’s diseases. Therefore, morphologically detailed alterations in mitochondria and 3D reconstruction of mitochondria are highly demanded research problems in the performance of clinical diagnosis. Nevertheless, manual mitochondria segmentation over 3D electron microscopy volumes is not a trivial task. This study proposes a two-stage cascaded CNN architecture to achieve automated 3D mitochondria segmentation, combining the merits of top-down and bottom-up approaches. For top-down approaches, the segmentation is conducted on objects’ localization so that the delineations of objects’ contours can be more precise. However, the combinations of 2D segmentation from the top-down approaches are inadequate to perform proper 3D segmentation without the information on connectivity among frames. On the other hand, the bottom-up approach finds coherent groups of pixels and takes the information of 3D connectivity into account in segmentation to avoid the drawbacks of the 2D top-down approach. However, many small areas that share similar pixel properties with mitochondria become false positives due to insufficient information on objects’ localization. In the proposed method, the detection of mitochondria is carried out with multi-slice fusion in the first stage, forming the segmentation cues. Subsequently, the second stage is to perform 3D CNN segmentation that learns the pixel properties and the information of 3D connectivity under the supervision of cues from the detection stage. Experimental results show that the proposed structure alleviates the problems in both the top-down and bottom-up approaches, which significantly accomplishes better performance in segmentation and expedites clinical analysis.

Identification and Classification of Lungs Focal Opacity Using CNN Segmentation and Optimal Feature Selection.

Lung cancer is one of the deadliest cancers around the world, with high mortality rate in comparison to other cancers. A lung cancer patient's survival probability in late stages is very low. However, if it can be detected early, the patient survival rate can be improved. Diagnosing lung cancer early is a complicated task due to having the visual similarity of lungs nodules with trachea, vessels, and other surrounding tissues that leads toward misclassification of lung nodules. Therefore, correct identification and classification of nodules is required. Previous studies have used noisy features, which makes results comprising. A predictive model has been proposed to accurately detect and classify the lung nodules to address this problem. In the proposed framework, at first, the semantic segmentation was performed to identify the nodules in images in the Lungs image database consortium (LIDC) dataset. Optimal features for classification include histogram oriented gradients (HOGs), local binary patterns (LBPs), and geometric features are extracted after segmentation of nodules. The results shown that support vector machines performed better in identifying the nodules than other classifiers, achieving the highest accuracy of 97.8% with sensitivity of 100%, specificity of 93%, and false positive rate of 6.7%.

CT Perfusion is All we Need: 4D CNN Segmentation of Penumbra and Core in Patients With Suspected Acute Ischemic Stroke

CT Perfusion is All we Need: 4D CNN Segmentation of Penumbra and Core in Patients With Suspected Acute Ischemic Stroke

3-D Visualization for Lung Covid-19 Infection Based on U-Net CNN Segmentation

Covid-19 is still the attention of researchers in medical image analysis. Following the initial respiratory diagnosis, a CT-scan examination will be performed. Segmentation of infection within the lung area is needed as the next step after the examination. In recent years, image segmentation has been carried out with the help of Deep Learning. U-Net CNN is one of the Deep Learning-based architectures widely used in medical image segmentation. Our research aims to support radiologists in visualizing Covid-19 infection in 3D based on CNN’s U-Net segmentation. It results in two types of visualization: 3D bitmap and 3D Mesh. 3D visualization can contribute to seeing the extent of infection and calculating the predicted percentage of Covid-19 infection volume in the patient’s lungs. The dataset for training model CNN is relatively small, consisting of 20 CT scans of Zenodo’s Covid-19 patients, divided into 17 patients, 2808 images (80%) for training, and three patients, 712 images (20%) for testing. The segmentation evaluation metrics used are DICE, Precision, and Accuracy, while the evaluation metrics for 3D volume are Relative Volume Difference (RVD) and Volumetric Similarity (VS). Finally, the prediction of the percentage of infection volume to the patient’s lung volume is given. The evaluation results were satisfactory, obtaining 95% DICE scores for lung image segmentation and 75% for infection segmentation. The predicted 3D visualization volume also scored higher than 90% for both lung volume and infection. Furthermore, for calculating the prediction of infection volume against lung volume, we achieved a maximum difference of 3% from the ground truth value.

Convolutional bi-directional learning and spatial enhanced attentions for lung tumor segmentation

Background and objective: Accurate lung tumor segmentation from computed tomography (CT) is complex due to variations in tumor sizes, shapes, patterns and growing locations. Learning semantic and spatial relations between different feature channels, image regions and positions is critical yet challenging. Methods: We propose a new segmentation method, PRCS, by learning and integrating multi-channel contextual relations, and spatial and position dependencies across image regions. Firstly, to extract contextual relationships between different deep image feature tensor channels, we propose a new convolutional bi-directional gated recurrent unit based module for forward and backward learning. Secondly, a novel cross-channel region-level attention mechanism is proposed to discriminate the contributions of different local regions and associated features in the global learning process. Finally, spatial and position dependencies are formulated by a new position-enhanced self-attention mechanism. The new attention can measure the diverse contributions of other positions to a target position and obtain an enhanced adaptive feature vector for the target position.Results: Our model outperformed seven state-of-the-art segmentation methods on both public and in-house lung tumor datasets in terms of spatial overlapping and shape similarity. Ablation study results proved the effectiveness of three technical innovations and generalization ability on different 3D CNN segmentation backbones. Conclusion: The proposed model enhanced the learning and propagation of contextual, spatial and position relations in 3D volumes, improving lung tumours’ segmentation performance with large variations and indistinct boundaries. PRCS provides an effective automated approach to support precision diagnosis and treatment planning of lung cancer.

Segmentation of Brain Tumors Using Three-Dimensional Convolutional Neural Network on MRI Images 3D MedImg-CNN

We consider the problem of fully automatic brain tumor segmentation in MR images containing glioblastomas. We propose a three Dimensional Convolutional Neural Network (3D MedImg-CNN) approach which achieves high performance while being extremely efficient, a balance that existing methods have struggled to achieve. Our 3D MedImg-CNN is formed directly on the raw image modalities and thus learn a characteristic representation directly from the data. We propose a new cascaded architecture with two pathways that each model normal details in tumors. Fully exploiting the convolutional nature of our model also allows us to segment a complete cerebral image in one minute. The performance of the proposed 3D MedImg-CNN with CNN segmentation method is computed using dice similarity coefficient (DSC). In experiments on the 2013, 2015 and 2017 BraTS challenges datasets; we unveil that our approach is among the most powerful methods in the literature, while also being very effective.

Fully Automatic Knee Joint Segmentation and Quantitative Analysis for Osteoarthritis from Magnetic Resonance (MR) Images Using a Deep Learning Model.

BackgroundWe aimed to develop and evaluate a deep learning-based method for fully automatic segmentation of knee joint MR imaging and quantitative computation of knee osteoarthritis (OA)-related imaging biomarkers.Material/MethodsThis retrospective study included 843 volumes of proton density-weighted fat suppression MR imaging. A convolutional neural network segmentation method with multiclass gradient harmonized Dice loss was trained and evaluated on 500 and 137 volumes, respectively. To assess potential morphologic biomarkers for OA, the volumes and thickness of cartilage and meniscus, and minimal joint space width (mJSW) were automatically computed and compared between 128 OA and 162 control data.ResultsThe CNN segmentation model produced reasonably high Dice coefficients, ranging from 0.948 to 0.974 for knee bone compartments, 0.717 to 0.809 for cartilage, and 0.846 for both lateral and medial menisci. The OA-related biomarkers computed from automatic knee segmentation achieved strong correlation with those from manual segmentation: average intraclass correlations of 0.916, 0.899, and 0.876 for volume and thickness of cartilage, meniscus, and mJSW, respectively. Volume and thickness measurements of cartilage and mJSW were strongly correlated with knee OA progression.ConclusionsWe present a fully automatic CNN-based knee segmentation system for fast and accurate evaluation of knee joint images, and OA-related biomarkers such as cartilage thickness and mJSW were reliably computed and visualized in 3D. The results show that the CNN model can serve as an assistant tool for radiologists and orthopedic surgeons in clinical practice and basic research.

Computational analysis of histological images from hematoxylin and eosin-stained oral epithelial dysplasia tissue sections

Oral epithelial dysplasia is a precancerous lesion that presents alterations in the shape and size of cell nuclei and can be graded as mild, moderate and severe. The conventional process for diagnosis of this lesion is complex, time-consuming and subject to errors. The use of digital systems in histological analysis can aid specialists to obtain data that allows a robust and fast investigation of the lesion. This work presents a method for dysplasia quantification in histopathological images of the oral cavity using machine learning models. The methodology includes the steps of nuclei segmentation, post-processing, feature extraction and classification. On the segmentation step, the Mask R-CNN neural network was trained using nuclei masks, where objects were detected. The post-processing step employed morphological operations to remove false positive and negative areas. Then, 23 morphological and non-morphological features such as area, orientation, solidity and entropy were computed and a polynomial classifier was employed to distinguish the images among the lesion’s grades. This approach was applied in a dataset with 296 regions of mice tongue images, where 9155 cell nuclei were identified and analysed . Metrics such as accuracy and area under the ROC curve were employed to evaluate the methodology by comparing it with the gold standard marked by specialists and other methods present in the literature. This work presents a novel study for the classification of automated grading of oral dysplasia lesions based on the association of CNN segmentation and polynomial algorithm. The segmentation step resulted in accuracies ranging from 88.92% to 90.35% and the classification step obtained area under the ROC curve ranging from 0.88 to 0.97. When compared to other algorithms present in the literature, our methods showed more relevant results, obtaining higher accuracy and AUC values. These values showed that the proposed methodology contributed to the state-of-the-art and can be used as a tool to aid pathologists with precise values for investigating dysplastic tissue lesions.

Time-Series Analysis and Prediction of Surface Deformation in the Jinchuan Mining Area, Gansu Province, by Using InSAR and CNN–PhLSTM Network

Surface deformation poses a great threat to the safety of Jinchuan mining area production activities. At present, the spatio-temporal evolution law and mechanism of surface deformation in the Jinchuan mining area are unclear, and it is difficult to obtain reliable prediction results using the existing spatio-temporal prediction methods due to the lack of model parameters or relevant data. To solve these problems, this study proposes a new unified convolutional neural network with peephole long short-term memory (CNN-PhLSTM). Small baseline subset interferometric synthetic aperture radar (SBAS-InSAR) technology was used to obtain the spatio-temporal evolution characteristics of surface deformation in the period of 2014–2021. Time series InSAR deformation data are merged into a unified network model in series with a time-distributed CNN segmentation and stacked PhLSTM. The InSAR measurement results are shown to be reliable by comparison and verification with the benchmark and InSAR results of different orbits. The proposed CNN-PhLSTM model was evaluated by mean absolute error and structural similarity (SSIM) evaluation indexes, and was compared with support vector regression (SVR), multilayer perceptron (MLP) and CNN-LSTM models. The results show three continuous subsidence areas, namely the Longshou, second western and third eastern mining areas. The cumulative surface deformation continued to increase from 2014 to 2021. Faults and lithology control the spatial distribution of surface deformation in the Jinchuan mining area. The prediction results demonstrate that the surface deformation range will continue to expand and that time-series surface deformation will show a slow deceleration trend in the next two years.

Developing an explainable deep learning boundary correction method by incorporating cascaded x-Dim models to improve segmentation defects in liver CT images

Deep learning methods achieved remarkable results in medical image analysis tasks but it has not yet been widely used by medical professionals. One of the main reasons for this restricted usage is the uncertainty of the reasons that influence the decision of the model. Explainable AI methods have been developed to improve the transparency, interpretability, and explainability of the black-box AI methods. The result of an explainable segmentation method will be more trusted by experts. In this study, we designed an explainable deep correction method by incorporating cascaded 1D and 2D models to refine the output of other models and provide reliable yet accurate results. We implemented a 2-step loop with a 1D local boundary validation model in the first step, and a 2D image patch segmentation model in the second step, to refine incorrect segmented regions slice-by-slice. The proposed method improved the result of the CNN segmentation models and achieved state-of-the-art results on 3D liver segmentation with the average dice coefficient of 98.27 on the Sliver07 dataset.

Evaluation of U-Net Based Architectures for Automatic Aortic Dissection Segmentation

Segmentation and reconstruction of arteries is important for a variety of medical and engineering fields, such as surgical planning and physiological modeling. However, manual methods can be laborious and subject to a high degree of human variability. In this work, we developed various convolutional neural network ( CNN ) architectures to segment Stanford type B aortic dissections ( TBADs ), characterized by a tear in the descending aortic wall creating a normal channel of blood flow called a true lumen and a pathologic channel within the wall called a false lumen. We introduced several variations to the two-dimensional ( 2D ) and three-dimensional (3 D ) U-Net, where small stacks of slices were inputted into the networks instead of individual slices or whole geometries. We compared these variations with a variety of CNN segmentation architectures and found that stacking the input data slices in the upward direction with 2D U-Net improved segmentation accuracy, as measured by the Dice similarity coefficient ( DC ) and point-by-point average distance ( AVD ), by more than 15\% . Our optimal architecture produced DC scores of 0.94, 0.88, and 0.90 and AVD values of 0.074, 0.22, and 0.11 in the whole aorta, true lumen, and false lumen, respectively. Altogether, the predicted reconstructions closely matched manual reconstructions.

Segmentation of Brain Tumors Using Three-Dimensional Convolutional Neural Network on MRI Images 3D MedImg-CNN

We consider the problem of fully automatic brain tumor segmentation in MR images containing glioblastomas. We propose a three Dimensional Convolutional Neural Network (3D MedImg-CNN) approach which achieves high performance while being extremely efficient, a balance that existing methods have struggled to achieve. Our 3D MedImg-CNN is formed directly on the raw image modalities and thus learn a characteristic representation directly from the data. We propose a new cascaded architecture with two pathways that each model normal details in tumors. Fully exploiting the convolutional nature of our model also allows us to segment a complete cerebral image in one minute. The performance of the proposed 3D MedImg-CNN with CNN segmentation method is computed using dice similarity coefficient (DSC). In experiments on the 2013, 2015 and 2017 BraTS challenges datasets; we unveil that our approach is among the most powerful methods in the literature, while also being very effective.

CNN TRANSFER LEARNING FOR DETECTION, COUNTING AND SEGMENTATION OF COCONUT PALMS FROM SATELLITE IMAGES

CNN TRANSFER LEARNING FOR DETECTION, COUNTING AND SEGMENTATION OF COCONUT PALMS FROM SATELLITE IMAGES

A Persistent Homology-Based Topological Loss Function for Multi-class CNN Segmentation of Cardiac MRI.

With respect to spatial overlap, CNN-based segmentation of short axis cardiovascular magnetic resonance (CMR) images has achieved a level of performance consistent with inter observer variation. However, conventional training procedures frequently depend on pixel-wise loss functions, limiting optimisation with respect to extended or global features. As a result, inferred segmentations can lack spatial coherence, including spurious connected components or holes. Such results are implausible, violating the anticipated topology of image segments, which is frequently known a priori. Addressing this challenge, published work has employed persistent homology, constructing topological loss functions for the evaluation of image segments against an explicit prior. Building a richer description of segmentation topology by considering all possible labels and label pairs, we extend these losses to the task of multi-class segmentation. These topological priors allow us to resolve all topological errors in a subset of 150 examples from the ACDC short axis CMR training data set, without sacrificing overlap performance.

Image Localized Style Transfer to Design Clothes Based on CNN and Interactive Segmentation.

In recent years, image style transfer has been greatly improved by using deep learning technology. However, when directly applied to clothing style transfer, the current methods cannot allow the users to self-control the local transfer position of an image, such as separating specific T-shirt or trousers from a figure, and cannot achieve the perfect preservation of clothing shape. Therefore, this paper proposes an interactive image localized style transfer method especially for clothes. We introduce additional image called outline image, which is extracted from content image by interactive algorithm. The interaction consists simply of dragging a rectangle around the desired clothing. Then, we introduce an outline loss function based on distance transform of the outline image, which can achieve the perfect preservation of clothing shape. In order to smooth and denoise the boundary region, total variation regularization is employed. The proposed method constrains that the new style is generated only in the desired clothing part rather than the whole image including background. Therefore, in our new generated images, the original clothing shape can be reserved perfectly. Experiment results show impressive generated clothing images and demonstrate that this is a good approach to design clothes.