Ecole d'ingénieur et centre de recherche en Sciences du numérique

Nonparametric scene parsing with deep convolutional features and dense alignment

Ma, Chih-Hao; Hsu, Chiou-Ting; Huet, Benoit

ICIP 2015, IEEE International Conference on Image Processing, September 27-30, 2015, Quebec, Canada

This paper addresses two key issues which concern the performance of nonparametric scene parsing: (1) the semantic quality of image retrieval; and (2) the accuracy in label transfer. First, because nonparametric methods annotate a query image through transferring labels from retrieved images, the task of image retrieval should find a set of "semantically similar" images to the query. Second, with the retrieval set, a good strategy should be developed to transfer semantic labels in pixel-level accuracy. In this paper, we focus on improving scene parsing accuracy in these two issues. We propose using the state-of-the-art deep convolutional features as image descriptors to improve the semantic quality of retrieved images. In addition, we include dense alignment into the Markov Random Field inference framework to transfer labels at pixel-level accuracy. Our experiments on the SIFT Flow dataset shows the improvement of the proposed approach over other nonparametric methods.

Document Bibtex

Titre:Nonparametric scene parsing with deep convolutional features and dense alignment
Mots Clés:scene parsing; object window; deep convolutional network; SIFT flow
Département:Data Science
Eurecom ref:4605
Copyright: © 2015 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Bibtex: @inproceedings{EURECOM+4605, year = {2015}, title = {{N}onparametric scene parsing with deep convolutional features and dense alignment}, author = {{M}a, {C}hih-{H}ao and {H}su, {C}hiou-{T}ing and {H}uet, {B}enoit }, booktitle = {{ICIP} 2015, {IEEE} {I}nternational {C}onference on {I}mage {P}rocessing, {S}eptember 27-30, 2015, {Q}uebec, {C}anada }, address = {{Q}u{\'e}bec, {CANADA}}, month = {09}, url = {} }
Voir aussi: