ISBI 2017

112 papers
The 2017 International Symposium on Biomedical Imaging (ISBI) featured 126 papers focused on the intersection of deep learning and clinical diagnostics. Prominent research themes included the use of high-dimensional radiomics to decode tumor phenotypes and the development of multi-faceted clinical prediction models for holistic patient assessment. Methodological advancements highlighted the integration of physical laws into Bayesian priors and the application of Markov Random Fields for spatial smoothing. Researchers also addressed practical challenges in data quality, such as modeling annotator reliability for noisy labels and image restoration using Euler's elastica inpainting. Notable applications presented this year included the automatic detection of aortic dissection in CT scans and the parcellation of the visual cortex on histological brain sections. Other significant work, contributed by prolific authors such as Tianming Liu, Yang Song, and Gustavo Carneiro, involved survival prediction using cascaded neural networks and cerebral vascular enhancement. These studies demonstrated a shift toward high-dimensional morphological profiling and complex-valued fMRI data analysis, reflecting the year's emphasis on sophisticated computational modeling in medical imaging.

Top Research Directions

Research directions most represented by ISBI 2017 papers.

Overcoming Handcrafted Feature Limitations in Biomedical Imaging 2
Native 3D Volumetric Feature Extraction 2
Continuous Domain Optimization for Subpixel Boundary Precision 1
Texture Segmentation Using Local Variance Metrics 1
Uncertainty Guided Spatial Refinement for Efficient Vision 1
Unsupervised Discovery of Latent Behavioral Primitives 1
Enforcing Monotonicity in Sequential Neural Predictions 1
Eliminating Manual Thresholds in Computer Vision 1
Enforcing Spatial Consistency with Markov Random Fields 1
Preventing Mesh Self Intersection and Element Inversion 1
Overcoming the Isointense Phase in Infant Neuroimaging 1
Kernel Based Non-linear Feature Mapping 1
Generalizing Neural Simulators to Unseen Physical Regimes 1
Scaling Kernel Methods with Random Fourier Features 1
Multi-Scale Time-Frequency Decomposition for Time Series Modeling 1

Top Authors

AuthorPapers
Tianming Liu6
Lei Guo4
Weidong Cai4
Yang Song4
Behzad Bozorgtabar3
Dinggang Shen3
Gustavo Carneiro3
Hugues Talbot3
Le Lu3
Michael Unser3
Roland Kwitt3
Xintao Hu3
Zongyuan Ge3
Babak Ehteshami Bejnordi2
Benjamin David Haeffele2
bjoern menze2
Daniel Schmitter2
Dong Ni2
Georgios Kaissis2
Jeffrey A Fessler2

Impactful Papers

Papers from ISBI 2017 that shaped research clusters.

Foundational

Influential

Papers

Title
Segmentation of Organs at Risk in thoracic CT images using a SharpMask architecture and Conditional Random Fields
Self supervised deep representation learning for fine-grained body part recognition
Simple domain adaptation for cross-dataset analyses of brain MRI data
Spatially variant PSF modeling in confocal macroscopy
Statistics on the space of trajectories for longitudinal data analysis
Structural connectome validation using pairwise classification
SurvivalNet: Predicting patient survival from diffusion weighted magnetic resonance images using cascaded fully convolutional and 3D Convolutional Neural Networks
SurvivalNet: Predicting patient survival from diffusion weighted magnetic resonance images using cascaded fully convolutional and 3D Convolutional Neural Networks.
Template-guided Functional Network Identification via Supervised Dictionary Learning
The importance of stain normalization in colorectal tissue classification with convolutional networks
Tree-loss function for training neural networks on weakly-labelled datasets
TumorNet: Lung nodule characterization using multi-view Convolutional Neural Network with Gaussian Process

« prev Page 3/3