Built-in Foreground/Background Prior for Weakly-Supervised Semantic Segmentation
Authors:
- Fatemehsadat SalehFSFatemehsadat Saleh
- Mohammad Sadegh Ali AkbarianMAMohammad Sadegh Ali Akbarian
- Mathieu SalzmannMSMathieu Salzmann
- Lars PeterssonLPLars Petersson
- Stephen GouldSGStephen Gould
- Jose M. AlvarezJAJose M. Alvarez
Preprint
Abstract
Pixel-level annotations are expensive and time consuming to obtain. Hence, weak supervision using only image tags could have a significant impact in semantic segmentation. Recently, CNN-based methods have proposed to fine-tune pre-trained networks using image tags. Without additional information, this leads to poor localization accuracy. This problem, however, was alleviated by making use of objectness priors to generate foreground/background masks. Unfortunately these priors either require training pixel-level annotations/bounding boxes, or still yield inaccurate object boundaries. Here, we propose a novel method to extract markedly more accurate masks from the pre-trained network itself, forgoing external objectness modules. This is accomplished using the activations of the higher-level convolutional layers, smoothed by a dense CRF. We demonstrate that our method, based on these masks and a weakly-supervised loss, outperforms the state-of-the-art tag-based weakly-supervised semantic segmentation techniques. Furthermore, we introduce a new form of inexpensive weak supervision yielding an additional accuracy boost.Keywords
All Related Versions
This publication has 5 references indexed in Scilit:
- STC: A Simple to Complex Framework for Weakly-Supervised Semantic SegmentationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2016
- Learning to segment with image-level annotationsPattern Recognition, 2016
- What’s the Point: Semantic Segmentation with Point Supervision2016
- DeepEdge: A multi-scale bifurcated deep network for top-down contour detection2015
- The Pascal Visual Object Classes Challenge: A RetrospectiveInternational Journal of Computer Vision, 2014