Seed, expand and constrain: Three principles for weakly-supervised image segmentation

Kolesnikov A, Lampert C. 2016. Seed, expand and constrain: Three principles for weakly-supervised image segmentation. ECCV: European Conference on Computer Vision, LNCS, vol. 9908, 695–711.

Download (ext.)

Conference Paper | Published | English

Scopus indexed
Department
Series Title
LNCS
Abstract
We introduce a new loss function for the weakly-supervised training of semantic image segmentation models based on three guiding principles: to seed with weak localization cues, to expand objects based on the information about which classes can occur in an image, and to constrain the segmentations to coincide with object boundaries. We show experimentally that training a deep convolutional neural network using the proposed loss function leads to substantially better segmentations than previous state-of-the-art methods on the challenging PASCAL VOC 2012 dataset. We furthermore give insight into the working mechanism of our method by a detailed experimental study that illustrates how the segmentation quality is affected by each term of the proposed loss function as well as their combinations.
Publishing Year
Date Published
2016-09-15
Publisher
Springer
Volume
9908
Page
695 - 711
Conference
ECCV: European Conference on Computer Vision
Conference Location
Amsterdam, The Netherlands
Conference Date
2016-10-11 – 2016-10-14
IST-REx-ID

Cite this

Kolesnikov A, Lampert C. Seed, expand and constrain: Three principles for weakly-supervised image segmentation. In: Vol 9908. Springer; 2016:695-711. doi:10.1007/978-3-319-46493-0_42
Kolesnikov, A., & Lampert, C. (2016). Seed, expand and constrain: Three principles for weakly-supervised image segmentation (Vol. 9908, pp. 695–711). Presented at the ECCV: European Conference on Computer Vision, Amsterdam, The Netherlands: Springer. https://doi.org/10.1007/978-3-319-46493-0_42
Kolesnikov, Alexander, and Christoph Lampert. “Seed, Expand and Constrain: Three Principles for Weakly-Supervised Image Segmentation,” 9908:695–711. Springer, 2016. https://doi.org/10.1007/978-3-319-46493-0_42.
A. Kolesnikov and C. Lampert, “Seed, expand and constrain: Three principles for weakly-supervised image segmentation,” presented at the ECCV: European Conference on Computer Vision, Amsterdam, The Netherlands, 2016, vol. 9908, pp. 695–711.
Kolesnikov A, Lampert C. 2016. Seed, expand and constrain: Three principles for weakly-supervised image segmentation. ECCV: European Conference on Computer Vision, LNCS, vol. 9908, 695–711.
Kolesnikov, Alexander, and Christoph Lampert. Seed, Expand and Constrain: Three Principles for Weakly-Supervised Image Segmentation. Vol. 9908, Springer, 2016, pp. 695–711, doi:10.1007/978-3-319-46493-0_42.
All files available under the following license(s):
Copyright Statement:
This Item is protected by copyright and/or related rights. [...]

Link(s) to Main File(s)
Access Level
OA Open Access

Export

Marked Publications

Open Data ISTA Research Explorer

Search this title in

Google Scholar