Lung CT Images Research Articles

Segmentation of Lung Ct Images using Cascaded Fully Convolutional Neural Networks

Interpretation of CT Lung images by the radiologist can be enhanced to a greater extent by automatic segmentation of nodules. The efficiency of this interpretation depends on the completeness and non-ambiguousness of the CT Lung images. Here, a fully automatic cascaded basis was proposed for CT Lung image segmentation. In this proposal a customized FCN was used feature extractions exploration from many visual scales and differentiate anatomy with a thick forecast map. Widespread experimental outcomes demonstrate that this technique can address the incompleteness in boundary and this technique can achieve best accuracy in segmentation of Lung CT Images when compared to other techniques which address the same area.

Radiation dosimetry changes in radiotherapy treatment plans for adult patients arising from the selection of the CT image reconstruction kernel.

Objective:The reconstruction kernel used for a CT scan strongly influences the image quality. This work investigates the changes in Hounsfield units (HUs) which can arise when altering the image reconstruction kernel for planning CT images and the associated changes in dose in the radiotherapy treatment plan if the treatment planning system (TPS) is not re-calibrated.Methods:Head and neck, prostate and lung CT images from four centres were used. For a specific scan, the base image was acquired using the original reconstruction kernel (used when the TPS was calibrated) and the treatment plan produced. The treatment plan was applied to all images from the other reconstruction kernels. Differences in dose-volume metrics for the planning target volume (PTV) and organs at risk (OARs) were noted and HU differences between images measured for air, soft tissue and bone.Results: HU change in soft tissue had the greatest influence on dose change. When within ±20 HU for soft tissue and ±50 HU for bone and air the dose change in the PTV and OAR was within ±0.5% and ±1% respectively.Conclusions: When imaging parameters were changed, if HU change was within ±20 HU for soft tissue and ±50 HU for bone and air, the change in the PTV and OAR doses was below 1%.Advances in knowledge:The degree of dose change in the treatment plan with HU change is demonstrated for current TPS algorithms. This adds to the limited evidence base for recommendations on HU tolerances as a tool for radiotherapy CT protocol optimization.

Development of Qualitative Model for Detection of Lung Cancer using Optimization

As of late, expectation of cancer at prior stages is mandatory to increase the opportunity of survival of the harassed. The most appalling sort is lung cancer, which is most common malady these days. So to dispose of it a detection framework is proposed. The objective of this paper is to investigate a practical segmentation algorithm with optimization system for therapeutic images to abridge the doctors' understanding of CT images. Recent medicinal imaging modalities produce enormous images that are incredibly terrible to examine physically. The outcomes of segmentation algorithm depend on the exactitude and intermingling time. In this paper, a qualitative detection model is proposed to partition the CT images of lung cancer. The detection framework shaped the obtained therapeutic images of lung CT images. To begin with, in pre-processing stage the median filter is utilized for noise reduction and smoothing. Later Otsu’s segmentation is applied to separate locale of enthusiasm from lung cancer images along with particle swarm optimization to get more accuracy and also for feature extraction LBP is connected. Here, the proposed model is framed by utilizing SVM technique for classification. Using MATLAB, simulation results are obtained for cancer detection system and these results are compared with other optimization techniques.

IoT with cloud based lung cancer diagnosis model using optimal support vector machine.

In the last decade, exponential growth of Internet of Things (IoT) and cloud computing takes the healthcare services to the next level. At the same time, lung cancer is identified as a dangerous disease which increases the global mortality rate annually. Presently, support vector machine (SVM) is the effective image classification tool especially in medical imaging. Feature selection and parameter optimization are the effective ways to improve the results of SVM and are conventionally resolved individually. This paper presents an optimal SVM for lung image classification where the parameters of SVM are optimized and feature selection takes place by modified grey wolf optimization algorithm combined with genetic algorithm (GWO-GA). The experimentation part takes place on three dimensions: test for parameter optimization, feature selection, and optimal SVM. For assessing the performance of the presented approach, a benchmark image database is employed which comprises of 50 low-dosage and stored lung CT images. The presented method exhibits its superior results on all the applied test images under several aspects. In addition, it achieves average classification accuracy of 93.54 which is significantly higher than the compared methods.

A new fast morphological geodesic active contour method for lung CT image segmentation

This work proposes a new adaptive approach to lung segmentation based on a non-parametric adaptive active contour method (ACM) without previous training using the new Fuzzy Border Detector, called Fast Morphological Geodesic Active Contour (FGAC). Performance was evaluated with 72 lung images of volunteers that were with fibrosis, chronic obstructive pulmonary disease or were healthy. A manual segmentation by a medical specialist was considered the gold standard. The mean time of segmentation of seventy-two the two lungs was 1.98 s. This was the best average time among all the other segmentation methods compared here: GVF (240 s), VFC(30 s), OPS Euclidean (5.86 s), SISDEP (4.90 s), and CRAD (2 s); thus showing its potential for real-time applications. The FGAC showed good results in all similarity metrics compared in this work like the Jaccard Index (92.73%), Dice coefficient (96.19%) and Matthew correlation coefficient (95.54%), and also achieved good results in sensitivity (99.21%), and accuracy (98.86%). The new approach showed quantitative indexes better than the traditional methods VFC, GVF, RHT mod, RHT multi. Moreover, we evaluated the proposed method against the supervised techniques OPS-Euclidean, SISDEP, CRISP and CRAD. Our approach achieved superior or equivalent results to these methods, however with a shorter convergence time. These results had good indexes suggesting that the proposed approach can be used to aid medical diagnosis in pulmonology.

A Deep Learning Technique For Lung Nodule Classification Based on False Positive Reduction

Cancer is of the major reasons of human death universally, one of the deadliest types of cancer is lung cancer, that causes the highest rate of the dead in both genders combined. Detecting lung cancer in early stage does not guarantee the survive of the patient’s life but it can reduce the mortality ratio by a high degree, early detection mainly includes screening unhealthy human’s lung using most valuable imaging modality which is CT scan. Classifying nodules in lung CT images adopting an automatic computer system become a necessary task due to a huge number of situations every day to help human expert’s in decision making procedure. Over the past few years, a numerous computer system is presented, each done a certain task such as detecting, segmenting, and classifying lung tumors using dissimilar algorithms. The objective of this study is to design an automated lung nodule classification system using two distinct deep learning architectures which are Network In Network (NIN) and standard Convolution Neural Network (CNN). The two models are trained and tested using 13,500 2D cubes around the nodule location that obtained from LUNA16 dataset, the database consists of 888 3D CT scans with annotation file determined a nodule position in every scan. The models are trained with a diverse cube size and hyperparameters in order to develop a high-performance structure for each model. The experimental results showed that best achieved scores for NIN are accuracy 90%, precision 99%, recall 68%, and false positive rate 0.06%, but for the typical CNN are accuracy 90%, precision 85%, recall 85%, and false positive rate 7.52%.

Lung cancer detection from CT image using improved profuse clustering and deep learning instantaneously trained neural networks

Automatic lung disease detection is a critical challenging task for researchers because of the noise signals getting included into creative signals amid the image capturing process which may corrupt the cancer image quality thusly bringing about the debased performance. So as to evade this, Lung cancer preprocessing has turned into an imperative stage with the key parts as edge detection, lung image resampling, lung image upgrade and image denoising for improving the nature of input image. Image Denoising is a critical pre-processing task preceding further preparing of the image like feature extraction, segmentation, surface examination, and so forth which elminates the noise whereas retaining the edges and additional complete features to the extent possible. This paper deals with improvement of the quality of lung image and diagnosis of lung cancer by reducing misclassification. The lung CT images are collected from Cancer imaging Archive (CIA) dataset, noise present in the images are eliminated by applying weighted mean histogram equalization approach which successfully removes noise from image, also enhancing the quality of the image, using improved profuse clustering technique (IPCT) for segmenting the affected region. Various spectral features are derived from the affected region. These are examined by applying deep learning instantaneously trained neural network for predicting lung cancer. Eventually, the system is examined by the efficiency of the system using MATLAB based simulation results. The system ensures that 98.42% of accuracy with minimum classification error 0.038.

Patch-based lung ventilation estimation using multi-layer supervoxels

Patch-based approaches have received substantial attention over the recent years in medical imaging. One of their potential applications may be to provide more anatomically consistent ventilation maps estimated on dynamic lung CT. An assessment of regional lung function may act as a guide for radiotherapy, ensuring a more accurate treatment plan. This in turn, could spare well-functioning parts of the lungs. We present a novel method for lung ventilation estimation from dynamic lung CT imaging, combining a supervoxel-based image representation with deformations estimated during deformable image registration, performed between peak breathing phases. For this we propose a method that tracks changes of the intensity of previously extracted supervoxels. For the evaluation of the method we calculate correlation of the estimated ventilation maps with static ventilation images acquired from hyperpolarized Xenon129 MRI. We also investigate the influence of different image registration methods used to estimate deformations between the peak breathing phases in the dynamic CT imaging. We show that our method performs favorably to other ventilation estimation methods commonly used in the field, independently of the image registration method applied to dynamic CT. Due to the patch-based approach of our method, it may be physiologically more consistent with lung anatomy than previous methods relying on voxel-wise relationships. In our method the ventilation is estimated for supervoxels, which tend to group spatially close voxels with similar intensity values. The proposed method was evaluated on a dataset consisting of three lung cancer patients undergoing radiotherapy treatment, and this resulted in a correlation of 0.485 with XeMRI ventilation images, compared with 0.393 for the intensity-based approach, 0.231 for the Jacobian-based method and 0.386 for the Hounsfield units averaging method, on average. Within the limitation of the small number of cases analyzed, results suggest that the presented technique may be advantageous for CT-based ventilation estimation. The results showing higher values of correlation of the proposed method demonstrate the potential of our method to more accurately mimic the lung physiology.

Analysis of Lung CT Images Based on Fisher Criterion and Genetic Optimization

Analysis of Lung CT Images Based on Fisher Criterion and Genetic Optimization

Fully Automated Coronal and Sagittal Chest Segmentation using Colour Features and Fuzzy C-Means Clustering in CT Images

In this article, a Combination of Fuzzy logic and color features based segmentation approach for parenchyma of lung from the Coronal and Sagittal Chest CT images is proposed. This approach employs a modified method of segmenting lung parenchyma which is considered as the Region of Interest (ROI) from the Coronal and Sagittal Chest CT images. The first step is the pre-processing of CT lung image in for the process of removing the noise and artefacts present in it. The border detection process is carried out as a Second step where all the regions including the tissue and lung parenchyma is separated by a border detection algorithm. Third step is the color formation process in which the image along with its border is formed in magenta color. Then, the color features extracted from the image border and are given as input to the Improved Fuzzy C Means clustering (IFCM) method to produce the lung Parenchyma. Experimental results reveal that the proposed methodology provides better segmentation results of 97.8 % accuracy in segmenting the parenchyma from Coronal and Sagittal Chest CT images.

HoTPiG: a novel graph-based 3-D image feature set and its applications to computer-assisted detection of cerebral aneurysms and lung nodules.

A novel image feature set named histogram of triangular paths in graph (HoTPiG) is presented. The purpose of this study is to evaluate the feasibility of the proposed HoTPiG feature set through two clinical computer-aided detection tasks: nodule detection in lung CT images and aneurysm detection in head MR angiography images. The HoTPiG feature set is calculated from an undirected graph structure derived from a binarized volume. The features are derived from a 3-D histogram in which each bin represents a triplet of shortest path distances between the target node and all possible node pairs near the target node. First, the vessel structure is extracted from CT/MR volumes. Then, a graph structure is extracted using an 18-neighbor rule. Using this graph, a HoTPiG feature vector is calculated at every foreground voxel. After explicit feature mapping with an exponential-χ2 kernel, each voxel is judged by a linear support vector machine classifier. The proposed method was evaluated using 300 CT and 300 MR datasets. The proposed method successfully detected lung nodules and cerebral aneurysms. The sensitivity was about 80% when the number of false positives was three per case for both applications. The HoTPiG image feature set was presented, and its high general versatility was shown through two medical lesion detection applications.

Comparison of CT ventilation imaging and hyperpolarised gas MRI: effects of breathing manoeuvre

Image registration of lung CT images acquired at different inflation levels has been proposed as a surrogate method to map lung ‘ventilation’. Prior to clinical use, it is important to understand how this technique compares with direct ventilation imaging modalities such as hyperpolarised gas MRI. However, variations in lung inflation level have been shown to affect regional ventilation distributions. Therefore, the aim of this study was to evaluate the impact of lung inflation levels when comparing CT ventilation imaging to ventilation from 3He-MRI.Seven asthma patients underwent breath-hold CT at total lung capacity (TLC) and functional residual capacity (FRC). 3He-MRI and a same-breath 1H-MRI were acquired at FRC+1L and TLC. Percentage ventilated volumes (%VVs) were calculated for FRC+1L and TLC 3He-MRI. TLC-CT and registered FRC-CT were used to compute a surrogate ventilation map from voxel-wise intensity differences in Hounsfield unit values, which was thresholded at the 10th and 20th percentiles. For direct comparison of CT and 3He-MRI ventilation, FRC+1L and TLC 3He-MRI were registered to TLC-CT indirectly via the corresponding same-breath 1H-MRI data. For 3He-MRI and CT ventilation comparison, Dice similarity coefficients (DSCs) between the binary segmentations were computed.The median (range) of %VVs for FRC+1L and TLC 3He-MRI were 90.5 (54.9–93.6) and 91.8 (67.8–96.2), respectively (p = 0.018). For MRI versus CT ventilation comparison, statistically significant improvements in DSCs were observed for TLC 3He MRI when compared with FRC+1L, with median (range) values of 0.93 (0.86–0.93) and 0.86 (0.68–0.92), respectively (p = 0.017), for the 10–100th percentile and 0.87 (0.83–0.88) and 0.81 (0.66–0.87), respectively (p = 0.027), for the 20–100th percentile.Correlation of CT ventilation imaging and hyperpolarised gas MRI is sensitive to lung inflation level. For ventilation maps derived from CT acquired at FRC and TLC, a higher correlation with gas ventilation MRI can be achieved if the MRI is acquired at TLC.

Characteristic Changes and 3D Virtual Measurement of Lung CT Image Parameters in the Drowning Rabbit Model.

To use virtual anatomy technique in the analysis of post-mortem characteristic changes of CT images in the experimental drowning rabbit model and the related parameters in 3D virtual model, so as to explore its application value in the diagnosis of drowning in forensic pathology. A model of drowning rabbits was established, with animal models of hemorrhagic shock and mechanical asphyxia as the controls. CT scan was performed on the experimental animals, and the differences in imaging features between the groups were compared by morphological reading of the tomographic images. CT data were imported into Mimics 14.0 software for 3D modeling. The CT values and lung volumes were calculated by the software, and the differences on CT values and lung volumes brought by different causes of death were analyzed. The CT images of lungs in the drowning group showed characteristic ground-glass opacity （diffuse and uniform density increase）. There were no obvious abnormalities in hemorrhagic shock group, and only a few similar changes were found in the mechanical asphyxia group. Compared with the controls, the CT values and the lung volumes in the drowning group were significantly increased P<0.05. Based on post-mortem lung imaging, the combination of CT value and lung volume changes can effectively reflect the virtual anatomical features in drowning, and provide a diagnostic basis for the forensic identification of drowning.

Mil based lung CT-image classification using CNN

Tuberculosis (TB) is diagnosed using clinical settings through which the specimen taken as of the patient is examined. The availability of Mycobacterium tuberculosis bacteria (MTB) in that specimen confirms the existence of TB. The other examinations strongly recommend that, TB in the diagnosis, one cannot confirm it. A simple skin test is the most generally utilized diagnostic equipment for TB, though blood tests are becoming more common. A small quantity of a substance termed PPD (Post-Partum Depression) tuberculin is injected below the skin of one’s inside forearm. Sometimes, false negative treatment may also results. Thus several upcoming techniques are introduced to cure the TB. It is based upon the CAD system that deals with the problem of TB detection on CXR. It is maintained by the training dataset. The drawback that occurs on the previous model is that its inherent features are uncertain. The novel 3 methods implemented in this research are i) Lung Segment, ii) Texture Feature Extraction as well as iii) Pixel classification. Along with these, miSVM+PEDD (multiple instances Support Vector Machine + Probability Estimation and Data Discarding) is used for the segmentation process. Thus the model should be evaluated by implementing some recent features of the Multiple Instance Learning (MIL). The Improved Algorithm is deployed by training a MIL classifier which builds from other machine learning (ML), Active Learning (AL) and 1-class classification approaches. Here, the uncertainty intrinsic to a MIL pixel classifier is diminished while minimizing the labeling exertions. The uses of AL are centered on the image and text categorization. The resulting lung likelihood map is post-processed to obtain binary segmentation. Here, the MIL is integrated with the AL Model. The results are analyzed by contrasting the proposed system with the other prevailing techniques to prove the dominance of the proposed one.

An Effective Diffeomorphic Model and Its Fast Multigrid Algorithm for Registration of Lung CT Images

Abstract Image registration is the process of aligning sets of similar, but different, intensity image functions to track changes between the images. In medical image problems involving lung images, variational registration models are a very powerful tool which can aid in effective treatment of various lung conditions and diseases. However, a common drawback of many variational models, such as the diffusion model and even optic flow models, is the lack of control of folding in the deformations leading to physically inaccurate transformations. For this reason, such models are generally not suitable for real life lung imaging problems where folding cannot occur. There are two approaches offering reliable solutions (though not necessarily accurate). The first approach is a parametric model such as the affine registration model, still widely used in many applications, but is unable to track local changes or yield accurate results. The second approach is to impose an extra constraint on the transformation of registration at the cost of increased non-linearity. An alternative to the second approach, achieving diffeomorphic transforms without adding any constraints, is an inverse consistent model such as by Christensen and Johnson (2001) from computing explicitly both the forward and inverse transforms. However, one must deal with the strong non-linearity in the formulation. In this paper we first propose a simplified inverse consistent model to avoid the inclusion of strong non-linearities and then a fast non-linear multigrid (NMG) technique to overcome the extra computational work required by the inverse consistent model. Experiments, performed on real medical CT images, show that our proposed inverse consistent model is robust to both parameter choice and non-folding in the transformations when compared with diffusion type models.

Computer aided diagnosis of drug sensitive pulmonary tuberculosis with cavities, consolidations and nodular manifestations on lung CT images

In this work, a computer aided diagnosis (CAD) system to improve the diagnostic accuracy and consistency in image interpretation of pulmonary tuberculosis is proposed. The lung fields are segmented using region growing and edge reconstruction algorithms. Texture features are extracted from the diseased regions manifested as consolidations, cavitations and nodular opacities. A wrapper approach that combines cuckoo search optimisation and one-against-all SVM classifier is used to select optimal feature subset. Cuckoo search algorithm is implemented first using entropy and second without using entropy measure. Training is done with the selected features using one-against-all (SVM) classifier. Among the 98 features extracted from the diseased regions, 47 features are selected with entropy measure giving 92.77% accuracy. When entropy measure is not used, 51 features are selected giving 91.89% accuracy. From the results, it is inferred that selecting appropriate features for training the classifier has an impact on the classifier performance.

Computer aided diagnosis of drug sensitive pulmonary tuberculosis with cavities, consolidations and nodular manifestations on lung CT images

In this work, a computer aided diagnosis (CAD) system to improve the diagnostic accuracy and consistency in image interpretation of pulmonary tuberculosis is proposed. The lung fields are segmented using region growing and edge reconstruction algorithms. Texture features are extracted from the diseased regions manifested as consolidations, cavitations and nodular opacities. A wrapper approach that combines cuckoo search optimisation and one-against-all SVM classifier is used to select optimal feature subset. Cuckoo search algorithm is implemented first using entropy and second without using entropy measure. Training is done with the selected features using one-against-all (SVM) classifier. Among the 98 features extracted from the diseased regions, 47 features are selected with entropy measure giving 92.77% accuracy. When entropy measure is not used, 51 features are selected giving 91.89% accuracy. From the results, it is inferred that selecting appropriate features for training the classifier has an impact on the classifier performance.

GUI-CAD Tool for Segmentation and Classification of Abnormalities in Lung CT Image

A vital necessity for clinical determination and treatment is an opportunity to prepare a procedure that is universally adaptable. Computer aided diagnosis (CAD) of various medical conditions has seen a tremendous growth in recent years. The frameworks combined with expanding capacity, the coliseum of CAD is touching new spaces. The goal of proposed work is to build an easy to understand multifunctional GUI Device for CAD that performs intelligent preparing of lung CT images. Functions implemented are to achieve region of interest (ROI) segmentation for nodule detection. The nodule extraction from ROI is implemented by morphological operations, reducing the complexity and making the system suitable for real-time applications. In addition, an interactive 3D viewer and performance measure tool that quantifies and measures the nodules is integrated. The results are validated through clinical expert. This serves as a foundation to determine, the decision of treatment and the prospect of recovery.

Isolated Pulmonary Nodules Characteristics Detection Based on CT Images

Pulmonary nodules are the main pathological changes of the lung. Malignant pulmonary nodules will be transformed into lung cancer, which is a serious threat to human health and life. Therefore, the detection of pulmonary nodules is of great significance to save lives. However, in the face of a large number of lung CT image sequences, doctors need to spend a lot of time and energy, and in the detection process will inevitably produce the problem of false detection and missed detection. Therefore, it is very necessary for computer-aided doctors to detect pulmonary nodules. It is difficult to segment pulmonary nodules accurately and recognize the characteristics of pulmonary nodules in CT images. A complete set of semi-automatic lung nodule extraction and feature identification system is established, which is in line with the doctor’s diagnosis process. A segmentation algorithm of pulmonary nodules based on regional statistical information is proposed to extract pulmonary nodules accurately. This is the first time that dynamic time warping algorithm is applied in the field of image processing, focusing on the lung nodule boundary. On this basis, the recursive graph visualization model is established to realize the visualization of boundary similarity. Finally, in order to accurately identify the characteristics of pulmonary nodules, a video similarity distance discrimination system is introduced to quantify the similarity between the nodules to be examined and the pulmonary nodules in the database. The experimental results show that the algorithm can accurately identify the normal shape, lobulated shape and lobulated shape of pulmonary nodules. The average processing speed is 0.58s/nodule. To some extent, it can reduce the misdiagnosis caused by experience and fatigue.

Tensor Voting Based Vascular Bifurcation Detection in CT Images of Lung

The accurate detection of vascular bifurcations is not helpful for pulmonary vascular disease diagnosis, but vital in (Computed Tomography) CT image analysis and processing of lung. We propose a tensor voting based method for vascular bifurcation detection in CT image of lung, which a vessel enhancement method is firstly proceeding to initially extract vessel structure, on which we perform ball voting with the pixels.