Training β-VAE by Aggregating a Learned Gaussian Posterior with a Decoupled Decoder

Article Computer Science, Artificial Intelligence

Automatic skull defect restoration and cranial implant generation for cranioplasty

Jianning Li et al.

Summary: The study presents a data-driven approach for the fast and fully automatic design of 3D printed patient-specific cranial implants. By training deep learning models on healthy skulls with synthetic defects, the models can be directly transferred to skulls with real defects, showing promise for clinical use.

MEDICAL IMAGE ANALYSIS (2021)

Add to Collection

Article Computer Science, Interdisciplinary Applications

AutoImplant 2020-First MICCAI Challenge on Automatic Cranial Implant Design

Jianning Li et al.

Summary: This paper provides a comprehensive overview of the MICCAI 2020 AutoImplant Challenge, summarizing the approaches and publications submitted, as well as the evaluation results. Based on feedback from collaborating neurosurgeons, open issues and post-challenge requirements for intra-operative use are discussed.

IEEE TRANSACTIONS ON MEDICAL IMAGING (2021)

Add to Collection

Article Computer Science, Artificial Intelligence

Use of Variational Autoencoders with Unsupervised Learning to Detect Incorrect Organ Segmentations at CT

Veit Sandfort et al.

Summary: A deep learning model was developed to detect incorrect organ segmentations in CT scans using variational autoencoders. Training three different 3D U-Nets and then training variational autoencoders to identify problematic segmentations proved effective in detecting errors. The study demonstrated the superiority of the deep learning method compared to traditional non-deep learning methods in identifying unusual or incorrect organ segmentations.

RADIOLOGY-ARTIFICIAL INTELLIGENCE (2021)

Add to Collection

Article Multidisciplinary Sciences

SkullBreak/SkullFix - Dataset for automatic cranial implant design and a benchmark for volumetric shape learning tasks

Oldrich Kodym et al.

Summary: This article introduces two complementary datasets, SkullBreak and SkullFix, for the development of data-driven solutions for cranial implant design. The datasets, which are adapted from a public head CT collection CQ500, are used for training and evaluation purposes. The datasets are not overlapping and serve as complements to each other, offering opportunities for evaluating volumetric shape learning algorithms.

DATA IN BRIEF (2021)

Add to Collection

Article Computer Science, Information Systems

Balancing Reconstruction Error and Kullback-Leibler Divergence in Variational Autoencoders

Andrea Asperti et al.

IEEE ACCESS (2020)

Add to Collection

Article Computer Science, Information Systems

Variational Autoencoder With Optimizing Gaussian Mixture Model Priors

Chunsheng Guo et al.

IEEE ACCESS (2020)

Add to Collection

Training β-VAE by Aggregating a Learned Gaussian Posterior with a Decoupled Decoder

Related references

Automatic skull defect restoration and cranial implant generation for cranioplasty

AutoImplant 2020-First MICCAI Challenge on Automatic Cranial Implant Design

Use of Variational Autoencoders with Unsupervised Learning to Detect Incorrect Organ Segmentations at CT

SkullBreak/SkullFix - Dataset for automatic cranial implant design and a benchmark for volumetric shape learning tasks

Balancing Reconstruction Error and Kullback-Leibler Divergence in Variational Autoencoders

Variational Autoencoder With Optimizing Gaussian Mixture Model Priors

Export Citation

Share Paper