Multi-atlas segmentation of the cartilage in knee MR images with sequential volume- and bone-mask-based registrations

Lee, Han Sang; Kim, Hyeun A.; Kim, Hyeonjin; Hong, Helen; Yoon, Young Cheol; Kim, Junmo

doi:10.1117/12.2216630

Cited by 1 publication

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As the cartilage’s shapes are thin elongated and varies due to pathological changes, pixel‐wise mapping based objective functions may not provide optimized segmentation performance. For example, although 2D UNet and 3D UNet improved the automatic cartilage and meniscus segmentation performance compared with the conventional models, their performance is still limited by the predefined traditional objective function. In our proposed model, the UNet performance is improved by the discriminator network which provides adversarial feedback and features for calculating the UNet loss during training of network.…”

Section: Discussionmentioning

confidence: 99%

“…The main reason for the improvement is the inclusion of batch normalization layers, which fix distributions of inputs at each hidden layer and remove the effect of the internal covariate shift . Moreover, the proposed model requires fewer preprocessing steps compared with semiautomatic or fully automatic atlas based or shape‐based methods, as it does not require registration to perform segmentation. Furthermore, the time required to perform segmentation is much less compared with these semi‐ or fully automatic methods.…”

Section: Discussionmentioning

confidence: 99%

“…In addition, these segmentations are affected by inter‐ and intrareader variability . To overcome these challenges, fully automatic cartilage segmentation methods were proposed where predefined features, intensity or shape‐based models, and multiple atlas were used for segmentation. However, the segmentation performances of these methods were not completely satisfactory due to the thin, elongated and irregular cartilage shape especially with pathological changes of cartilage structure.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Automated cartilage and meniscus segmentation of knee MRI with conditional generative adversarial networks

Gaj

Yang

Nakamura

et al. 2019

Magnetic Resonance in Med

View full text Add to dashboard Cite

Purpose Fully automatic tissue segmentation is an essential step to translate quantitative MRI techniques to clinical setting. The goal of this study was to develop a novel approach based on the generative adversarial networks for fully automatic segmentation of knee cartilage and meniscus. Theory and Methods Defining proper loss function for semantic segmentation to enforce the learning of multiscale spatial constraints in an end‐to‐end training process is an open problem. In this work, we have used the conditional generative adversarial networks to improve segmentation performance of convolutional neural network, such as UNet alone by overcoming the problems caused by pixel‐wise mapping based objective functions, and to capture cartilage features during the training of the network. Furthermore, the Dice coefficient and cross entropy losses were incorporated to the loss functions to improve the model performance. The model was trained and tested on 176, 3D DESS (double‐echo steady‐state) knee images from the Osteoarthritis Initiative data set. Results The proposed model provided excellent segmentation performance for cartilages with Dice coefficients ranging from 0.84 in patellar cartilage to 0.91 in lateral tibial cartilage, with an average Dice coefficient of 0.88. For meniscus segmentation, the model achieves 0.89 Dice coefficient for lateral meniscus and 0.87 Dice coefficient for medial meniscus. The results are superior to previously published automatic cartilage and meniscus segmentation methods based on deep learning models such as convolutional neural network. Conclusion The proposed UNet‐conditional generative adversarial networks based model demonstrated a fully automated segmentation method with high accuracy for knee cartilage and meniscus.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation