Abstract

Semantic segmentation, by which an image is decomposed into regions with their respective semantic labels, is often the first step towards image understanding. Existing research on this regard is mainly performed under two conditions: the fully-supervised setting that relies on a set of images with pixel-level labels and the weakly-supervised one that uses only image-level labels. In both cases, the labeling task is time-consuming and laborious, and thus training data are always limited. In practice, there are voluminous on-line images, which unfortunately often have only incomplete image-level labels (tags) but would otherwise be potentially useful for a learning-based algorithm. Only limited efforts have been attempted on using such coarsely and incompletely labelled data for semantic segmentation. This paper proposes a new approach to semantic segmentation of a set of partially-labelled images, using a formulation considering information from multiple visual similar images. Experiments on several popular datasets, with comparison with existing methods, demonstrate evident performance improvement of the proposed approach.

Original languageEnglish (US)
Title of host publication2016 IEEE Winter Conference on Applications of Computer Vision, WACV 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509006410
DOIs
StatePublished - May 23 2016
EventIEEE Winter Conference on Applications of Computer Vision, WACV 2016 - Lake Placid, United States
Duration: Mar 7 2016Mar 10 2016

Other

OtherIEEE Winter Conference on Applications of Computer Vision, WACV 2016
CountryUnited States
CityLake Placid
Period3/7/163/10/16

Fingerprint

Labels
Semantics
Image understanding
Labeling
Pixels
Experiments

ASJC Scopus subject areas

  • Computer Science Applications
  • Computer Vision and Pattern Recognition

Cite this

Tian, Q., & Li, B. (2016). Simultaneous semantic segmentation of a set of partially labeled images. In 2016 IEEE Winter Conference on Applications of Computer Vision, WACV 2016 [7477639] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/WACV.2016.7477639

Simultaneous semantic segmentation of a set of partially labeled images. / Tian, Qiongjie; Li, Baoxin.

2016 IEEE Winter Conference on Applications of Computer Vision, WACV 2016. Institute of Electrical and Electronics Engineers Inc., 2016. 7477639.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Tian, Q & Li, B 2016, Simultaneous semantic segmentation of a set of partially labeled images. in 2016 IEEE Winter Conference on Applications of Computer Vision, WACV 2016., 7477639, Institute of Electrical and Electronics Engineers Inc., IEEE Winter Conference on Applications of Computer Vision, WACV 2016, Lake Placid, United States, 3/7/16. https://doi.org/10.1109/WACV.2016.7477639
Tian Q, Li B. Simultaneous semantic segmentation of a set of partially labeled images. In 2016 IEEE Winter Conference on Applications of Computer Vision, WACV 2016. Institute of Electrical and Electronics Engineers Inc. 2016. 7477639 https://doi.org/10.1109/WACV.2016.7477639
Tian, Qiongjie ; Li, Baoxin. / Simultaneous semantic segmentation of a set of partially labeled images. 2016 IEEE Winter Conference on Applications of Computer Vision, WACV 2016. Institute of Electrical and Electronics Engineers Inc., 2016.
@inproceedings{377ffef5f08646fd91a6960bf3e43e9b,
title = "Simultaneous semantic segmentation of a set of partially labeled images",
abstract = "Semantic segmentation, by which an image is decomposed into regions with their respective semantic labels, is often the first step towards image understanding. Existing research on this regard is mainly performed under two conditions: the fully-supervised setting that relies on a set of images with pixel-level labels and the weakly-supervised one that uses only image-level labels. In both cases, the labeling task is time-consuming and laborious, and thus training data are always limited. In practice, there are voluminous on-line images, which unfortunately often have only incomplete image-level labels (tags) but would otherwise be potentially useful for a learning-based algorithm. Only limited efforts have been attempted on using such coarsely and incompletely labelled data for semantic segmentation. This paper proposes a new approach to semantic segmentation of a set of partially-labelled images, using a formulation considering information from multiple visual similar images. Experiments on several popular datasets, with comparison with existing methods, demonstrate evident performance improvement of the proposed approach.",
author = "Qiongjie Tian and Baoxin Li",
year = "2016",
month = "5",
day = "23",
doi = "10.1109/WACV.2016.7477639",
language = "English (US)",
booktitle = "2016 IEEE Winter Conference on Applications of Computer Vision, WACV 2016",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - Simultaneous semantic segmentation of a set of partially labeled images

AU - Tian, Qiongjie

AU - Li, Baoxin

PY - 2016/5/23

Y1 - 2016/5/23

N2 - Semantic segmentation, by which an image is decomposed into regions with their respective semantic labels, is often the first step towards image understanding. Existing research on this regard is mainly performed under two conditions: the fully-supervised setting that relies on a set of images with pixel-level labels and the weakly-supervised one that uses only image-level labels. In both cases, the labeling task is time-consuming and laborious, and thus training data are always limited. In practice, there are voluminous on-line images, which unfortunately often have only incomplete image-level labels (tags) but would otherwise be potentially useful for a learning-based algorithm. Only limited efforts have been attempted on using such coarsely and incompletely labelled data for semantic segmentation. This paper proposes a new approach to semantic segmentation of a set of partially-labelled images, using a formulation considering information from multiple visual similar images. Experiments on several popular datasets, with comparison with existing methods, demonstrate evident performance improvement of the proposed approach.

AB - Semantic segmentation, by which an image is decomposed into regions with their respective semantic labels, is often the first step towards image understanding. Existing research on this regard is mainly performed under two conditions: the fully-supervised setting that relies on a set of images with pixel-level labels and the weakly-supervised one that uses only image-level labels. In both cases, the labeling task is time-consuming and laborious, and thus training data are always limited. In practice, there are voluminous on-line images, which unfortunately often have only incomplete image-level labels (tags) but would otherwise be potentially useful for a learning-based algorithm. Only limited efforts have been attempted on using such coarsely and incompletely labelled data for semantic segmentation. This paper proposes a new approach to semantic segmentation of a set of partially-labelled images, using a formulation considering information from multiple visual similar images. Experiments on several popular datasets, with comparison with existing methods, demonstrate evident performance improvement of the proposed approach.

UR - http://www.scopus.com/inward/record.url?scp=84977648880&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84977648880&partnerID=8YFLogxK

U2 - 10.1109/WACV.2016.7477639

DO - 10.1109/WACV.2016.7477639

M3 - Conference contribution

BT - 2016 IEEE Winter Conference on Applications of Computer Vision, WACV 2016

PB - Institute of Electrical and Electronics Engineers Inc.

ER -