Pixel Objectness
This addresses the need for robust foreground segmentation in computer vision applications like image retrieval and retargeting, though it is incremental as it builds on existing structured prediction and deep learning methods.
The paper tackles the problem of generating pixel-level foreground object segmentations from single images, even for unseen object categories, by training a deep fully convolutional network with mixed image-level and boundary-level annotations. It achieves substantial improvements in state-of-the-art on ImageNet and MIT Object Discovery datasets and generalizes well to over 1 million images with unseen categories.
We propose an end-to-end learning framework for generating foreground object segmentations. Given a single novel image, our approach produces pixel-level masks for all "object-like" regions---even for object categories never seen during training. We formulate the task as a structured prediction problem of assigning foreground/background labels to all pixels, implemented using a deep fully convolutional network. Key to our idea is training with a mix of image-level object category examples together with relatively few images with boundary-level annotations. Our method substantially improves the state-of-the-art on foreground segmentation for ImageNet and MIT Object Discovery datasets. Furthermore, on over 1 million images, we show that it generalizes well to segment object categories unseen in the foreground maps used for training. Finally, we demonstrate how our approach benefits image retrieval and image retargeting, both of which flourish when given our high-quality foreground maps.