Skip to Main content Skip to Navigation
Conference papers

Scale-Adaptive Forest Training via an Efficient Feature Sampling Scheme

Loic Peter 1 Olivier Pauly 1 Pierre Chatelain 1, 2 Diana Mateus 1 Nassir Navab 1, 3, 4
2 Lagadic - Visual servoing in robotics, computer vision, and augmented reality
CRISAM - Inria Sophia Antipolis - Méditerranée , Inria Rennes – Bretagne Atlantique , IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE
4 CAMP - Computer Aided Medical Procedures
Laboratory for Computational Sensing and Robotics
Abstract : In the context of forest-based segmentation of medical data, modeling the visual appearance around a voxel requires the choice of the scale at which contextual information is extracted, which is of crucial im- portance for the final segmentation performance. Building on Haar-like visual features, we introduce a simple yet effective modification of the for- est training which automatically infers the most informative scale at each stage of the procedure. Instead of the standard uniform sampling during node split optimization, our approach draws candidate features sequen- tially in a fine-to-coarse fashion. While being very easy to implement, this alternative is free of additional parameters, has the same computa- tional cost as a standard training and shows consistent improvements on three medical segmentation datasets with very different properties.
Complete list of metadata

Cited literature [15 references]  Display  Hide  Download

https://hal.inria.fr/hal-01241978
Contributor : Eric Marchand <>
Submitted on : Friday, December 18, 2015 - 2:44:32 PM
Last modification on : Friday, July 10, 2020 - 4:25:31 PM
Long-term archiving on: : Saturday, March 19, 2016 - 10:20:20 AM

File

peter2015miccai.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01241978, version 1

Citation

Loic Peter, Olivier Pauly, Pierre Chatelain, Diana Mateus, Nassir Navab. Scale-Adaptive Forest Training via an Efficient Feature Sampling Scheme. Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015, Oct 2015, Munich, Germany. ⟨hal-01241978⟩

Share

Metrics

Record views

1130

Files downloads

483