PET-CT images have been widely used in clinical practice for radiotherapy treatment planning of the radiotherapy. Many existing segmentation approaches only work for a single imaging modality, which suffer from the low spatial resolution in PET or low contrast in CT. In this work we propose a novel method for the co-segmentation of the tumor in both PET and CT images, which makes use of advantages from each modality: the functionality information from PET and the anatomical structure information from CT. The approach formulates the segmentation problem as a minimization problem of a Markov Random Field (MRF) model, which encodes the information from both modalities. The optimization is solved using a graph-cut based method. Two sub-graphs are constructed for the segmentation of the PET and the CT images, respectively. To achieve consistent results in two modalities, an adaptive context cost is enforced by adding context arcs between the two subgraphs. An optimal solution can be obtained by solving a single maximum flow problem, which leads to simultaneous segmentation of the tumor volumes in both modalities. The proposed algorithm was validated in robust delineation of lung tumors on 23 PET-CT datasets and two head-and-neck cancer subjects. Both qualitative and quantitative results show significant improvement compared to the graph cut methods solely using PET or CT.
Tumor segmentation in PET and CT images is notoriously challenging due to the low spatial resolution in PET and low contrast in CT images. In this paper, we have proposed a general framework to use both PET and CT images simultaneously for tumor segmentation. Our method utilizes the strength of each imaging modality: the superior contrast of PET and the superior spatial resolution of CT. We formulate this problem as a Markov Random Field (MRF) based segmentation of the image pair with a regularized term that penalizes the segmentation difference between PET and CT. Our method simulates the clinical practice of delineating tumor simultaneously using both PET and CT, and is able to concurrently segment tumor from both modalities, achieving globally optimal solutions in low-order polynomial time by a single maximum flow computation. The method was evaluated on clinically relevant tumor segmentation problems. The results showed that our method can effectively make use of both PET and CT image information, yielding segmentation accuracy of 0.85 in Dice similarity coefficient and the average median hausdorff distance (HD) of 6.4 mm, which is 10 % (resp., 16 %) improvement compared to the graph cuts method solely using the PET (resp., CT) images.
Dempster-Shafer evidence theory is a primary methodology for multi-source information fusion because it is good at dealing with uncertain information. This theory provides a Dempster’s rule of combination to synthesize multiple evidences from various information sources. However, in some cases, counter-intuitive results may be obtained based on that combination rule. Numerous new or improved methods have been proposed to suppress these counter-intuitive results based on perspectives, such as minimizing the information loss or deviation. Inspired by evolutionary game theory, this paper considers a biological and evolutionary perspective to study the combination of evidences. An evolutionary combination rule (ECR) is proposed to help find the most biologically supported proposition in a multi-evidence system. Within the proposed ECR, we develop a Jaccard matrix game (JMG) to formalize the interaction between propositions in evidences, and utilize the replicator dynamics to mimick the evolution of propositions. Experimental results show that the proposed ECR can effectively suppress the counter-intuitive behaviors appeared in typical paradoxes of evidence theory, compared with many existing methods. Properties of the ECR, such as solution’s stability and convergence, have been mathematically proved as well.
In this paper, we propose a new Multi-Criteria Decision-Making method (MCDM) which is rank reversal free. We call it the SPOTIS method standing for Stable Preference Ordering Towards Ideal Solution method. Our method is exempt of rank reversal because the preference ordering established from the score matrix of the MCDM problem under consideration does not require the relative comparisons between alternatives, but only comparisons with respect to the ideal solution chosen by the MCDM system designer after transforming the incomplete original MCDM problem into a well-defined one thanks to the specification of the min and max bounds of each criterion involved in the problem.
Purpose: The purpose of this work is twofold: First, to characterize the artifacts occurring in helical 4D-CT imaging; second, to propose a method that can automatically identify the artifacts in 4D-CT images. The authors have designed a process that can automatically identify the artifacts in 4D-CT images, which may be invaluable in quantifying the quality of 4D-CT images and reducing the artifacts from the reconstructed images on a large dataset. Methods: Given two adjacent stacks obtained from the same respiration phase, the authors determine if there are artifacts between them. The proposed method uses a "bridge" stack strategy to connect the two stacks. Using normalized cross correlation convolution ͑NCCC͒, the two stacks are mapped to the bridge stack and the best matching positions can be located. Using this position information, the authors can then determine if there are artifacts between the two stacks. By combining the matching positions with NCCC values, the performance can be improved. Results: To validate the method, three expert observers independently labeled over 600 stacks on five patients. The results confirmed that high performance was obtained using the proposed method. The average sensitivity was about 0.87 and the average specificity was 0.82. The proposed method also outperformed the method of using respiratory signal ͑sensitivity increased from 0.50 to 0.87 and specificity increased from 0.70 to 0.82͒. Conclusions: This study shows that the spatial artifacts during 4D-CT imaging are characterized and can be located automatically by the proposed method. The method is relatively simple but effective. It provides a way to evaluate the artifacts more objectively and accurately. The reported approach has promising potential for automatically identifying the types and frequency of artifacts on large scale 4D-CT image dataset.
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.