Sparse group composition for robust left ventricular epicardium segmentation

Wang, Bing; Gu, Xiaomeng; Fan, Chonghao; Xie, Hongzhi; Zhang, Shuyang; Tian, Xuedong; Gu, Lixu

doi:10.1016/j.compmedimag.2015.06.003

Cited by 5 publications

(4 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The advantage of the serial mode is that it can directly reserve the patient information in the model and compute the optimal parameters of the model by iteratively applying the expectation step and the maximization step of the expectation‐maximization algorithms proposed in Ref. [30].…”

Section: Methodsmentioning

confidence: 99%

“…Inspired by the sparse‐shape‐composition (SSC) model to model shape priors, we introduced the spatial relationship of multiple objects into the shape composition model to propose the SGC model . In this work, the SGC model is used to obtain the prior shape of a lung with an ML adhesion tumor.…”

Section: Methodsmentioning

confidence: 99%

“…With the intensive investigation of sparse representation, the sparse constraints have been used to model shape priors, which explicitly model spatially contiguous gross errors (non-Gaussian errors) in input shapes with sparse vectors and assume that these gross errors are sparse with respect to the given shape information. [26][27][28][29][30] However, this assumption does not hold when very large tumors are attached to lung boundaries because the non-Gaussian errors caused by these tumors are no longer sparse relative to the input lung shapes.…”

Section: Introductionmentioning

confidence: 99%

“…In the past, numerous methods have been proposed to mainly segment on small lung masses attached to lung boundaries (e.g., juxta‐pleural nodules) including region‐growing or watershed followed by morphological processing, supervised segmentation models based on feature extraction and support vector machines or neural networks, and energy minimization models with prior knowledge . With the intensive investigation of sparse representation, the sparse constraints have been used to model shape priors, which explicitly model spatially contiguous gross errors (non‐Gaussian errors) in input shapes with sparse vectors and assume that these gross errors are sparse with respect to the given shape information . However, this assumption does not hold when very large tumors are attached to lung boundaries because the non‐Gaussian errors caused by these tumors are no longer sparse relative to the input lung shapes.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Spline curve deformation model with prior shapes for identifying adhesion boundaries between large lung tumors and tissues around lungs in CT images

et al. 2020

Self Cite

View full text Add to dashboard Cite

Purpose Automated segmentation of lung tumors attached to anatomic structures such as the chest wall or mediastinum remains a technical challenge because of the similar Hounsfield units of these structures. To address this challenge, we propose herein a spline curve deformation model that combines prior shapes to correct large spatially contiguous errors (LSCEs) in input shapes derived from image‐appearance cues.The model is then used to identify the adhesion boundaries between large lung tumors and tissue around the lungs. Methods The deformation of the whole curve is driven by the transformation of the control points (CPs) of the spline curve, which are influenced by external and internal forces. The external force drives the model to fit the positions of the non‐LSCEs of the input shapes while the internal force ensures the local similarity of the displacements of the neighboring CPs. The proposed model corrects the gross errors in the lung input shape caused by large lung tumors, where the initial lung shape for the model is inferred from the training shapes by shape group‐based sparse prior information and the input lung shape is inferred by adaptive‐thresholding‐based segmentation followed by morphological refinement. Results The accuracy of the proposed model is verified by applying it to images of lungs with either moderate large‐sized (ML) tumors or giant large‐sized (GL) tumors. The quantitative results in terms of the averages of the dice similarity coefficient (DSC) and the Jaccard similarity index (SI) are 0.982 ± 0.006 and 0.965 ± 0.012 for segmentation of lungs adhered by ML tumors, and 0.952 ± 0.048 and 0.926 ± 0.059 for segmentation of lungs adhered by GL tumors, which give 0.943 ± 0.021 and 0.897 ± 0.041 for segmentation of the ML tumors, and 0.907 ± 0.057 and 0.888 ± 0.091 for segmentation of the GL tumors, respectively. In addition, the bidirectional Hausdorff distances are 5.7 ± 1.4 and 11.3 ± 2.5 mm for segmentation of lungs with ML and GL tumors, respectively. Conclusions When combined with prior shapes, the proposed spline curve deformation can deal with large spatially consecutive errors in object shapes obtained from image‐appearance information. We verified this method by applying it to the segmentation of lungs with large tumors adhered to the tissue around the lungs and the large tumors. Both the qualitative and quantitative results are more accurate and repeatable than results obtained with current state‐of‐the‐art techniques.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%