Dense and Low-Rank Gaussian CRFs Using Deep Embeddings - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2017

Dense and Low-Rank Gaussian CRFs Using Deep Embeddings

Abstract

In this work we introduce a structured prediction model that endows the Deep Gaussian Conditional Random Field (G-CRF) with a densely connected graph structure. We keep memory and computational complexity under control by expressing the pairwise interactions as inner products of low-dimensional, learnable embeddings. The G-CRF system matrix is therefore low-rank, allowing us to solve the resulting system in a few milliseconds on the GPU by using conjugate gradient. As in G-CRF, inference is exact, the unary and pairwise terms are jointly trained end-to-end by using analytic expressions for the gradients, while we also develop even faster, Potts-type variants of our embeddings. We show that the learned embeddings capture pixel-to-pixel affinities in a task-specific manner, while our approach achieves state of the art results on three challenging benchmarks, namely semantic segmentation, human part segmentation, and saliency estimation. Our implementation is fully GPU based, built on top of the Caffe library, and is available at https://github.com/ siddharthachandra/gcrf-v2.0.
Fichier principal
Vignette du fichier
chandra-iccv-2017.pdf (3.06 Mo) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01646293 , version 1 (23-11-2017)

Identifiers

  • HAL Id : hal-01646293 , version 1

Cite

Siddhartha Chandra, Nicolas Usunier, Iasonas Kokkinos. Dense and Low-Rank Gaussian CRFs Using Deep Embeddings. ICCV 2017 - International Conference on Computer Vision, Sep 2017, Venice, Italy. ⟨hal-01646293⟩
433 View
269 Download

Share

Gmail Facebook X LinkedIn More