One-Shot Segmentation in Clutter

Claudio Michaelis, Matthias Bethge, Alexander Ecker
Proceedings of the 35th International Conference on Machine Learning, PMLR 80:3549-3558, 2018.

Abstract

We tackle the problem of one-shot segmentation: finding and segmenting a previously unseen object in a cluttered scene based on a single instruction example. We propose a novel dataset, which we call cluttered Omniglot. Using a baseline architecture combining a Siamese embedding for detection with a U-net for segmentation we show that increasing levels of clutter make the task progressively harder. Using oracle models with access to various amounts of ground-truth information, we evaluate different aspects of the problem and show that in this kind of visual search task, detection and segmentation are two intertwined problems, the solution to each of which helps solving the other. We therefore introduce MaskNet, an improved model that attends to multiple candidate locations, generates segmentation proposals to mask out background clutter and selects among the segmented objects. Our findings suggest that such image recognition models based on an iterative refinement of object detection and foreground segmentation may provide a way to deal with highly cluttered scenes.

Cite this Paper


BibTeX
@InProceedings{pmlr-v80-michaelis18a, title = {One-Shot Segmentation in Clutter}, author = {Michaelis, Claudio and Bethge, Matthias and Ecker, Alexander}, booktitle = {Proceedings of the 35th International Conference on Machine Learning}, pages = {3549--3558}, year = {2018}, editor = {Dy, Jennifer and Krause, Andreas}, volume = {80}, series = {Proceedings of Machine Learning Research}, month = {10--15 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v80/michaelis18a/michaelis18a.pdf}, url = {https://proceedings.mlr.press/v80/michaelis18a.html}, abstract = {We tackle the problem of one-shot segmentation: finding and segmenting a previously unseen object in a cluttered scene based on a single instruction example. We propose a novel dataset, which we call cluttered Omniglot. Using a baseline architecture combining a Siamese embedding for detection with a U-net for segmentation we show that increasing levels of clutter make the task progressively harder. Using oracle models with access to various amounts of ground-truth information, we evaluate different aspects of the problem and show that in this kind of visual search task, detection and segmentation are two intertwined problems, the solution to each of which helps solving the other. We therefore introduce MaskNet, an improved model that attends to multiple candidate locations, generates segmentation proposals to mask out background clutter and selects among the segmented objects. Our findings suggest that such image recognition models based on an iterative refinement of object detection and foreground segmentation may provide a way to deal with highly cluttered scenes.} }
Endnote
%0 Conference Paper %T One-Shot Segmentation in Clutter %A Claudio Michaelis %A Matthias Bethge %A Alexander Ecker %B Proceedings of the 35th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2018 %E Jennifer Dy %E Andreas Krause %F pmlr-v80-michaelis18a %I PMLR %P 3549--3558 %U https://proceedings.mlr.press/v80/michaelis18a.html %V 80 %X We tackle the problem of one-shot segmentation: finding and segmenting a previously unseen object in a cluttered scene based on a single instruction example. We propose a novel dataset, which we call cluttered Omniglot. Using a baseline architecture combining a Siamese embedding for detection with a U-net for segmentation we show that increasing levels of clutter make the task progressively harder. Using oracle models with access to various amounts of ground-truth information, we evaluate different aspects of the problem and show that in this kind of visual search task, detection and segmentation are two intertwined problems, the solution to each of which helps solving the other. We therefore introduce MaskNet, an improved model that attends to multiple candidate locations, generates segmentation proposals to mask out background clutter and selects among the segmented objects. Our findings suggest that such image recognition models based on an iterative refinement of object detection and foreground segmentation may provide a way to deal with highly cluttered scenes.
APA
Michaelis, C., Bethge, M. & Ecker, A.. (2018). One-Shot Segmentation in Clutter. Proceedings of the 35th International Conference on Machine Learning, in Proceedings of Machine Learning Research 80:3549-3558 Available from https://proceedings.mlr.press/v80/michaelis18a.html.

Related Material