Deep Active Contours

TL;DR

This paper introduces a deep learning-based interactive boundary extraction method that combines CNN predictions with active contours, enabling efficient and accurate segmentation on medical and non-medical datasets.

Contribution

It presents a novel integration of CNN-based vector field prediction with active contour models for interactive segmentation, optimized for small hardware.

Findings

Effective boundary extraction on medical and non-medical datasets

Requires minimal computational resources

Compatible with small graphics cards

Abstract

We propose a method for interactive boundary extraction which combines a deep, patch-based representation with an active contour framework. We train a class-specific convolutional neural network which predicts a vector pointing from the respective point on the evolving contour towards the closest point on the boundary of the object of interest. These predictions form a vector field which is then used for evolving the contour by the Sobolev active contour framework proposed by Sundaramoorthi et al. The resulting interactive segmentation method is very efficient in terms of required computational resources and can even be trained on comparatively small graphics cards. We evaluate the potential of the proposed method on both medical and non-medical challenge data sets, such as the STACOM data set and the PASCAL VOC 2012 data set.