Visual Geometry Group - University of Oxford

Varun Gulshan, Carsten Rother, Antonio Criminisi, Andrew Blake and Andrew Zisserman

Overview

In this work [3] we introduce a new shape constraint for interactive image segmentation. It is an extension of Veksler's [1] star-convexity prior, in two ways: from a single star to multiple stars and from Euclidean rays to Geodesic paths. Global minima of the energy function are obtained subject to these new constraints. We also introduce Geodesic Forests, which exploit the structure of shortest paths in implementing the extended constraints. The star-convexity prior is used here in an interactive setting and this is demonstrated in a practical system. The system is evaluated by means of a robot user to measure the amount of interaction required in a precise way. We also introduce a new and harder dataset which augments the existing Grabcut dataset with images and ground truth taken from the PASCAL VOC segmentation challenge.

Code

Click here to get the code. Code includes user-interface implementing all the methods in the paper and the simulated user.

Dataset

Click here to get the dataset used for evaluation. The dataset consists of 151 images: 49 images are taken from the GrabCut dataset, 99 from the PASCAL VOC dataset and 3 images from the Alpha matting dataset.

Evaluation

Evaluation is done using a robot user [2] that simulates user interaction, and accuracy vs. effort plots are obtained. Refer to the paper for the plots and evaluation metrics.