Instance-sensitive Fully Convolutional Networks

TL;DR

This paper introduces an instance-sensitive FCN that generates multiple score maps to propose object instances for segmentation, leveraging local coherence rather than high-dimensional mask layers.

Contribution

It develops a novel FCN architecture that produces instance-sensitive score maps for improved object instance segmentation proposals.

Findings

Achieves competitive results on PASCAL VOC and MS COCO datasets.

Does not rely on high-dimensional mask layers, unlike previous methods.

Utilizes local image coherence for instance estimation.

Abstract

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.