Automated foveola localization in retinal 3D-OCT images using structural support vector machine prediction

Authors:
Yu-Ying Liu;Hiroshi Ishikawa;Mei Chen;Gadi Wollstein;Joel S. Schuman;James M. Rehg
Affiliations:
College of Computing, Georgia Institute of Technology, Atlanta, GA;UPMC Eye Center, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA, Department of Bioengineering, University of Pittsburgh, Pittsburgh, PA;Intel Science and Technology Center on Embedded Computing, Pittsburgh, PA;UPMC Eye Center, University of Pittsburgh School of Medicine, Pittsburgh, PA;UPMC Eye Center, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA, Department of Bioengineering, University of Pittsburgh, Pittsburgh, PA;College of Computing, Georgia Institute of Technology, Atlanta, GA
Venue:
MICCAI'12 Proceedings of the 15th international conference on Medical Image Computing and Computer-Assisted Intervention - Volume Part I
Year:
2012

Citing 5
Cited 0

Support vector machine learning for interdependent and structured output spaces

ICML '04 Proceedings of the twenty-first international conference on Machine learning
Active Scheduling of Organ Detection and Segmentation in Whole-Body Medical Images

MICCAI '08 Proceedings of the 11th international conference on Medical Image Computing and Computer-Assisted Intervention - Part I
Learning to Localize Objects with Structured Output Regression

ECCV '08 Proceedings of the 10th European Conference on Computer Vision: Part I
Cutting-plane training of structural SVMs

Machine Learning
Recursive coarse-to-fine localization for fast object detection

ECCV'10 Proceedings of the 11th European conference on Computer vision: Part VI

Quantified Score

Hi-index	0.00

Visualization

Abstract

We develop an automated method to determine the foveola location in macular 3D-OCT images in either healthy or pathological conditions. Structural Support Vector Machine (S-SVM) is trained to directly predict the location of the foveola, such that the score at the ground truth position is higher than that at any other position by a margin scaling with the associated localization loss. This S-SVM formulation directly minimizes the empirical risk of localization error, and makes efficient use of all available training data. It deals with the localization problem in a more principled way compared to the conventional binary classifier learning that uses zero-one loss and random sampling of negative examples. A total of 170 scans were collected for the experiment. Our method localized 95.1% of testing scans within the anatomical area of the foveola. Our experimental results show that the proposed method can effectively identify the location of the foveola, facilitating diagnosis around this important landmark.