Fast and Exact: ADMM-Based Discriminative Shape Segmentation with Loopy Part Models

Abstract : In this work we use loopy part models to segment ensembles of organs in medical images. Each organ's shape is represented as a cyclic graph, while shape consistency is enforced through inter-shape connections. Our contributions are two-fold: firstly, we use an efficient decomposition-coordination algorithm to solve the resulting optimization problems: we decompose the model's graph into a set of open, chain-structured, graphs each of which is efficiently optimized using Dynamic Program-ming with Generalized Distance Transforms. We use the Alternating Direction Method of Multipliers (ADMM) to fix the potential inconsistencies of the individual solutions and show that ADMM yields substantially faster convergence than plain Dual Decomposition-based methods. Secondly, we employ structured prediction to encompass loss functions that better reflect the performance criteria used in medical image segmentation. By using the mean contour distance (MCD) as a structured loss during training, we obtain clear test-time performance gains. We demonstrate the merits of exact and efficient inference with rich, structured models in a large X-Ray image segmentation benchmark, where we obtain systematic improvements over the current state-of-the-art.
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [45 references]  Display  Hide  Download
Contributor : Stavros Tsogkas <>
Submitted on : Monday, January 26, 2015 - 4:07:39 AM
Last modification on : Thursday, February 7, 2019 - 5:29:35 PM
Document(s) archivé(s) le : Monday, April 27, 2015 - 10:10:49 AM


Files produced by the author(s)




Haithem Boussaid, Iasonas Kokkinos. Fast and Exact: ADMM-Based Discriminative Shape Segmentation with Loopy Part Models. IEEE Conference in Computer Vision and Pattern Recognition, Jun 2014, Greater Columbus Convention Center in Columbus, Ohio., United States. pp.4058 - 4065, ⟨10.1109/CVPR.2014.517⟩. ⟨hal-01109287⟩



Record views


Files downloads