Instance-sensitive Fully Convolutional Networks
This work addresses the limitation of FCNs in distinguishing object instances for semantic segmentation, which is important for applications like autonomous driving and robotics, though it is incremental as it builds on existing FCN and DeepMask approaches.
The paper tackles the problem of instance-level segmentation by developing fully convolutional networks that generate instance-sensitive score maps, enabling the proposal of instance-level segment candidates. The method achieves competitive results on PASCAL VOC and MS COCO datasets.
Fully convolutional networks (FCNs) have been proven very successful for semantic segmentation, but the FCN outputs are unaware of object instances. In this paper, we develop FCNs that are capable of proposing instance-level segment candidates. In contrast to the previous FCN that generates one score map, our FCN is designed to compute a small set of instance-sensitive score maps, each of which is the outcome of a pixel-wise classifier of a relative position to instances. On top of these instance-sensitive score maps, a simple assembling module is able to output instance candidate at each position. In contrast to the recent DeepMask method for segmenting instances, our method does not have any high-dimensional layer related to the mask resolution, but instead exploits image local coherence for estimating instances. We present competitive results of instance segment proposal on both PASCAL VOC and MS COCO.