Exploiting saliency for object segmentation from image level labels
This work addresses the challenge of reducing annotation costs for semantic segmentation, making it more accessible for applications like autonomous driving or medical imaging, though it is incremental as it builds on existing weakly supervised methods.
The paper tackles the problem of training a pixel-wise semantic segmentation network using only image-level labels, which is ill-posed due to co-occurrences, by incorporating a saliency model as prior knowledge to recover 80% of fully supervised performance, setting a new state of the art in weakly supervised segmentation.
There have been remarkable improvements in the semantic labelling task in the recent years. However, the state of the art methods rely on large-scale pixel-level annotations. This paper studies the problem of training a pixel-wise semantic labeller network from image-level annotations of the present object classes. Recently, it has been shown that high quality seeds indicating discriminative object regions can be obtained from image-level labels. Without additional information, obtaining the full extent of the object is an inherently ill-posed problem due to co-occurrences. We propose using a saliency model as additional information and hereby exploit prior knowledge on the object extent and image statistics. We show how to combine both information sources in order to recover 80% of the fully supervised performance - which is the new state of the art in weakly supervised training for pixel-wise semantic labelling. The code is available at https://goo.gl/KygSeb.