Abstract:Distinguishing malignant lung nodules from benign nodules is an important aspect of lung cancer diagnosis. In this paper, we propose an automatic method to classify lung nodules into four different types, i.e. well-circumscribed, juxta-vascular, juxta-pleural and pleural-tail. Additionally, since the morphology of lung nodules forms a continuum between the different types, our proposed method is superior to previous methods that classify single nodules into a single type. First, a weighted similarity network i… Show more
“…The first three ones obtain better results among all control approaches. CPMw was used to identify the overlapping nodules in [10] in order to improve the classification; however, without concerning such issue, the proposed method still achieves better result, with 3% more of nodules correctly classified. Overall, it is apparent that our proposed method results in the best performance, suggesting its promising ability for lung nodule classification.…”
Section: B Resultsmentioning
confidence: 93%
“…The comparisons are conducted among the following methods: the proposed method, the weighed clique percolation method (CPMw) upon SVM probability estimates [10], SVM classification upon SIFT descriptor [10], linear discriminant analysis (LDA) upon SIFT descriptor [15], principle component analysis (PCA) upon SIFT descriptor [15] and the standard k-NN upon SIFT descriptor. Fig.10 shows the average classification rate across all training percentages (10%-90%) for each of these methods.…”
Section: B Resultsmentioning
confidence: 99%
“…However, the above techniques mostly emphasize on the nodule structures, restricting their ability to differentiate the intra-type nodules due to overlapping feature spaces of the nodules from different categories [10]. In addition, lung nodule is so complicated with various anatomical structures that those techniques would introduce large degree of unnecessary feature variations when concerning too many image details.…”
In this paper, a feature-based imaging classification method is presented to classify the lung nodules in low dose computed tomography (LDCT) slides into four categories: wellcircumscribed, vascularized, juxta-pleural and pleural-tail. The proposed method focuses on the feature design, which describes both lung nodule and surrounding context information, and contains two main stages: (1) superpixel labeling, which labels the pixels into foreground and background based on an image patch division approach, (2) context curve calculation, which transfers the superpixel labeling result into feature vector. While the first stage preprocesses the image, extracting the major context anatomical structures for each type of nodules, the context curve provides a discriminative description for intra-and inter-type nodules. The evaluation is conducted on a publicly available dataset and the results indicate the promising performance of the proposed method on lung nodule classification.
“…The first three ones obtain better results among all control approaches. CPMw was used to identify the overlapping nodules in [10] in order to improve the classification; however, without concerning such issue, the proposed method still achieves better result, with 3% more of nodules correctly classified. Overall, it is apparent that our proposed method results in the best performance, suggesting its promising ability for lung nodule classification.…”
Section: B Resultsmentioning
confidence: 93%
“…The comparisons are conducted among the following methods: the proposed method, the weighed clique percolation method (CPMw) upon SVM probability estimates [10], SVM classification upon SIFT descriptor [10], linear discriminant analysis (LDA) upon SIFT descriptor [15], principle component analysis (PCA) upon SIFT descriptor [15] and the standard k-NN upon SIFT descriptor. Fig.10 shows the average classification rate across all training percentages (10%-90%) for each of these methods.…”
Section: B Resultsmentioning
confidence: 99%
“…However, the above techniques mostly emphasize on the nodule structures, restricting their ability to differentiate the intra-type nodules due to overlapping feature spaces of the nodules from different categories [10]. In addition, lung nodule is so complicated with various anatomical structures that those techniques would introduce large degree of unnecessary feature variations when concerning too many image details.…”
In this paper, a feature-based imaging classification method is presented to classify the lung nodules in low dose computed tomography (LDCT) slides into four categories: wellcircumscribed, vascularized, juxta-pleural and pleural-tail. The proposed method focuses on the feature design, which describes both lung nodule and surrounding context information, and contains two main stages: (1) superpixel labeling, which labels the pixels into foreground and background based on an image patch division approach, (2) context curve calculation, which transfers the superpixel labeling result into feature vector. While the first stage preprocesses the image, extracting the major context anatomical structures for each type of nodules, the context curve provides a discriminative description for intra-and inter-type nodules. The evaluation is conducted on a publicly available dataset and the results indicate the promising performance of the proposed method on lung nodule classification.
“…To classify the lung nodules SVM classifier and PLSA (Probabilistic Latent Semantic Analysis) are proposed for the nodule patches and context patches in classification process [1], [4] and SVM classifier is utilized for the probabilistic estimation of the lung nodule [4].…”
This paper proposes a survey on the classification techniques of lung nodules. We have the different classifications about the nodules in the lungs. It contains the different methods of classification, segmentation and detection techniques. Malignant cell presented in the lungs named , nodules are classified for the treatment processes. Thresholding and Robust segmentation techniques are used in the segmentation process and the feature set is used for classification. Low Dose CT(Computed Tomography) images are applied. This survey has the information about the efficient techniques which are all used for the nodule classification. In these days lung cancer is the dangerous dead disease in the world, So we need to have the knowledge of that cancer. In starting stages the micro nodules are then formed into a cancer cell. Among the cancer affected population about 20% of the people are dead due to lung cancer. If nodules are found in a starting stage, we can be extend the lifetime of the patient. The main process of this paper involves with the nodule classification and segmentation process of the lung nodules. Here we taken the different procedures involved with nodule detections. CT is the most appropriate imaging technique to obtain anatomical information about lung nodules and the surrounding structures. Here we taken the Low Dose CT(LDCT) images for operations. This paper has the various approaches of the nodule classification. In this survey different techniques are presented which are used for detection and classification of the nodules in the lungs. By differentiating the nodules from the anatomical parts of the lungs, the nodules are identified.
“…Finally, classifier is constructed by a fusing method [14]. Zhang et al use a supervised learning method to find four probability values that belongs to each type [15]. Then, a weighed Clique Percolation method is implemented to discover the overlapping of lung nodules that belong to different type.…”
This paper proposes a novel lung nodule classification method for low-dose CT images. The method includes two stages. First, Local Difference Pattern (LDP) is proposed to encode the feature representation, which is extracted by comparing intensity difference along circular regions centered at the lung nodule. Then, the single-center classifier is trained based on LDP. Due to the diversity of feature distribution for different class, the training images are further clustered into multiple cores and the multicenter classifier is constructed. The two classifiers are combined to make the final decision. Experimental results on public dataset show the superior performance of LDP and the combined classifier.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.